https://easy-explain-app.vercel.app/ 2026-11-01 1.0 https://easy-explain-app.vercel.app/playlists 2026-11-01 0.9 https://easy-explain-app.vercel.app/playlist/PLKvb3M11cl9Q 2026-10-24 0.8 https://easy-explain-app.vercel.app/playlist/PLNBMQNhZSP5c 2026-10-29 0.8 https://easy-explain-app.vercel.app/playlist/PLVcQsK5b0mpA 2026-11-01 0.8 https://easy-explain-app.vercel.app/playlist/PLNb7oagc_zsg 2026-10-25 0.8 https://easy-explain-app.vercel.app/playlist/PLcuJu1K-wP1E 2026-10-30 0.8 https://easy-explain-app.vercel.app/playlist/PLGcd9hFBmpQQ 2026-11-01 0.8 https://easy-explain-app.vercel.app/lesson/6ABpSaJbSVI 2026-11-01 0.7 https://i.ytimg.com/vi/6ABpSaJbSVI/hqdefault.jpg https://i.ytimg.com/vi/6ABpSaJbSVI/hqdefault.jpg Why Your Backtest Order Can't Fill on the Same Bar Your order can't fill on the bar that triggered it. Here is the one-line rule that stops lookahead in Kestrel's simulated venue. https://www.youtube.com/embed/6ABpSaJbSVI 64 2026-11-01T10:38:34.116Z yes backtesting lookahead bias algorithmic trading C++20 trading engine fill model simulated venue slippage tick size NSE Indian markets quant developer event-driven Kestrel systems programming https://easy-explain-app.vercel.app/lesson/D3p-y0PyBxc 2026-11-01 0.7 https://i.ytimg.com/vi/D3p-y0PyBxc/hqdefault.jpg https://i.ytimg.com/vi/D3p-y0PyBxc/hqdefault.jpg Never Call a Handler Under a Lock One mutex mistake can freeze an entire trading engine. Here is how Kestrel's EventBus drain avoids it. https://www.youtube.com/embed/D3p-y0PyBxc 54 2026-11-01T10:38:29.020Z yes c++ c++20 mutex deadlock concurrency event bus systems programming kestrel trading engine lock std::deque swap software design event driven multithreading https://easy-explain-app.vercel.app/lesson/RyzvvNwa2jo 2026-11-01 0.7 https://i.ytimg.com/vi/RyzvvNwa2jo/hqdefault.jpg https://i.ytimg.com/vi/RyzvvNwa2jo/hqdefault.jpg The Bug That Made Every Strategy One Bar Late One constant shifted every backtest by one bar. Here is how Kestrel's store read bar end times wrong, and the three-line fix. https://www.youtube.com/embed/RyzvvNwa2jo 55 2026-11-01T10:38:23.729Z yes backtesting bug off by one bar lookahead bias C++20 algorithmic trading trading platform code Kestrel event-driven engine quant developer bar resampling time series bugs software engineering FYERS API NSE systems programming https://easy-explain-app.vercel.app/lesson/gVxbTL4y5xo 2026-11-01 0.7 https://i.ytimg.com/vi/gVxbTL4y5xo/hqdefault.jpg https://i.ytimg.com/vi/gVxbTL4y5xo/hqdefault.jpg A Day in Paper Trading: Running a C++ Trading Stack, and 15 Lessons What does it actually take to run an algorithmic trading stack for one day? We walk through Kestrel's daily paper-trading routine, from the morning script to the ordered shutdown, then recap the design lessons of the whole 26-episode course. https://www.youtube.com/embed/gVxbTL4y5xo 753 2026-11-01T10:38:18.859Z yes C++20 algorithmic trading paper trading trading system design event-driven architecture FYERS API NSE Indian markets systems programming software architecture lessons learned Unix sockets fail closed write-ahead log Kestrel quant developer https://easy-explain-app.vercel.app/lesson/OlKum2tFHqA 2026-11-01 0.7 https://i.ytimg.com/vi/OlKum2tFHqA/hqdefault.jpg https://i.ytimg.com/vi/OlKum2tFHqA/hqdefault.jpg Ten Layers of Live-Trading Safety: Every Switch Before a Real Order How many switches stand between a strategy and a real broker order? In Kestrel, the answer is ten, and none of them trusts the others. https://www.youtube.com/embed/OlKum2tFHqA 709 2026-11-01T10:38:03.598Z yes live trading safety defence in depth C++20 algorithmic trading FYERS API trading system design risk management software reconciliation kill switch environment variables Swiss cheese model systems programming Indian markets Kestrel order management system https://easy-explain-app.vercel.app/lesson/iN16nRT9xf0 2026-11-01 0.7 https://i.ytimg.com/vi/iN16nRT9xf0/hqdefault.jpg https://i.ytimg.com/vi/iN16nRT9xf0/hqdefault.jpg Kestrel Order Service: Sessions, Shadow Books, Crash Recovery What happens to your orders when a trading process dies? Kestrel's answer: the order service owns the truth, and the strategy keeps only a shadow. https://www.youtube.com/embed/iN16nRT9xf0 660 2026-11-01T10:37:47.021Z yes C++20 algorithmic trading order management system crash recovery idempotency write-ahead journal Unix sockets systems programming NSE FYERS API trading engine design shadow book fail closed software architecture Kestrel https://easy-explain-app.vercel.app/lesson/s3vZM-5t8ZA 2026-11-01 0.7 https://i.ytimg.com/vi/s3vZM-5t8ZA/hqdefault.jpg https://i.ytimg.com/vi/s3vZM-5t8ZA/hqdefault.jpg Safe Order Placement: When the Broker Doesn't Answer What should a trading engine do when an order request times out and nobody knows whether it reached the exchange? https://www.youtube.com/embed/s3vZM-5t8ZA 864 2026-11-01T10:37:34.181Z yes safe order placement write-ahead ledger idempotency order tag FYERS API C++20 algorithmic trading trading engine design uncertain outcomes crash recovery Kestrel Indian markets systems programming double fill retry safety https://easy-explain-app.vercel.app/lesson/REmSuZBv9Kc 2026-11-01 0.7 https://i.ytimg.com/vi/REmSuZBv9Kc/hqdefault.jpg https://i.ytimg.com/vi/REmSuZBv9Kc/hqdefault.jpg Kestrel Broker Plumbing: Daily Login, Secret Redaction, Rate Limiters What happens before a trading system sends its first request? Kestrel has to log in every morning, keep secrets out of every log, and stay under a broker's rate limits across several processes. https://www.youtube.com/embed/REmSuZBv9Kc 671 2026-11-01T10:37:17.255Z yes C++20 algorithmic trading FYERS API rate limiter sliding window OAuth login CSRF state secret redaction flock atomic file write systems programming Kestrel NSE Indian markets symbol master trading infrastructure https://easy-explain-app.vercel.app/lesson/OwN6ZhVZjKE 2026-11-01 0.7 https://i.ytimg.com/vi/OwN6ZhVZjKE/hqdefault.jpg https://i.ytimg.com/vi/OwN6ZhVZjKE/hqdefault.jpg One Connection, Many Traders: Inside Kestrel's Market-Data Service How does one process serve many strategies from a single broker connection without duplicating work or letting one slow client hurt the rest? https://www.youtube.com/embed/OwN6ZhVZjKE 724 2026-11-01T10:37:01.704Z yes C++20 algorithmic trading market data websocket fan-out Unix socket subscription refcount backpressure REST fallback FYERS API NSE event-driven architecture systems programming poll loop Kestrel bar builder https://easy-explain-app.vercel.app/lesson/qkZd9fvyPtU 2026-11-01 0.7 https://i.ytimg.com/vi/qkZd9fvyPtU/hqdefault.jpg https://i.ytimg.com/vi/qkZd9fvyPtU/hqdefault.jpg Decoding the FYERS Binary Market-Data Protocol, Byte by Byte Seventeen bytes arrive on a websocket. Can you turn them into a price safely? We decode the FYERS HSM binary market-data protocol byte by byte in C++20, inside the Kestrel trading platform. https://www.youtube.com/embed/qkZd9fvyPtU 775 2026-11-01T10:36:45.451Z yes FYERS FYERS API HSM protocol binary protocol websocket C++20 market data algorithmic trading Kestrel parsing binary data big-endian NSE Indian markets systems programming bounds checking Inside Kestrel https://easy-explain-app.vercel.app/lesson/vBpYOEogxZ8 2026-10-31 0.7 https://i.ytimg.com/vi/vBpYOEogxZ8/hqdefault.jpg https://i.ytimg.com/vi/vBpYOEogxZ8/hqdefault.jpg Kestrel: A Bounded HTTP Server, dlopen libcurl, and a Hand-Built WebSocket What if your trading binary could use libcurl without ever linking it, and speak WebSocket without a library? https://www.youtube.com/embed/vBpYOEogxZ8 776 2026-10-31T09:39:03.087Z 4 yes C++20 WebSocket RFC 6455 libcurl dlopen HTTP server systems programming networking XOR mask Kestrel algorithmic trading thread pool TLS Manim low level programming https://easy-explain-app.vercel.app/lesson/NGBnBN60qBg 2026-10-31 0.7 https://i.ytimg.com/vi/NGBnBN60qBg/hqdefault.jpg https://i.ytimg.com/vi/NGBnBN60qBg/hqdefault.jpg Wire Framing and Loopback Security: Inside Kestrel's Local Sockets Four bytes of length, a streaming parser that rejects a 4 GB lie instantly, and a browser tab that hits three walls. https://www.youtube.com/embed/NGBnBN60qBg 845 2026-10-31T09:38:46.486Z 6 yes C++20 length-prefixed framing Unix domain sockets loopback security constant-time compare DNS rebinding CSRF Origin header Host header timing attack non-blocking sockets MSG_NOSIGNAL algorithmic trading platform systems programming Kestrel socket hardening https://easy-explain-app.vercel.app/lesson/41xVtq8a8Iw 2026-10-31 0.7 https://i.ytimg.com/vi/41xVtq8a8Iw/hqdefault.jpg https://i.ytimg.com/vi/41xVtq8a8Iw/hqdefault.jpg Why Kestrel Runs as Many Processes: The Live Architecture One engine, many processes: why Kestrel splits paper and live trading into kestrel_md, kestrel_orders and kestrel_trader. https://www.youtube.com/embed/41xVtq8a8Iw 721 2026-10-31T09:38:26.891Z 12 yes C++20 systems programming algorithmic trading software architecture multi-process design Unix sockets signal handling sigtimedwait rate limiting FYERS API Indian markets NSE process isolation Kestrel trading platform https://easy-explain-app.vercel.app/lesson/W8ETv5s93Ig 2026-10-31 0.7 https://i.ytimg.com/vi/W8ETv5s93Ig/hqdefault.jpg https://i.ytimg.com/vi/W8ETv5s93Ig/hqdefault.jpg Golden Tests, Fakes and the Run Archive: Proving a Trading Engine Unchanged How do you refactor a trading engine and prove that not a single byte of behaviour changed? https://www.youtube.com/embed/W8ETv5s93Ig 847 2026-10-31T09:38:10.052Z 15 yes golden tests snapshot testing doctest C++20 unit testing C++ deterministic testing SHA-256 fake transports run archive producer consumer mkdir atomic algorithmic trading platform backtesting engine Kestrel refactoring safely ctest systems programming https://easy-explain-app.vercel.app/lesson/Myr5kcD89y8 2026-10-31 0.7 https://i.ytimg.com/vi/Myr5kcD89y8/hqdefault.jpg https://i.ytimg.com/vi/Myr5kcD89y8/hqdefault.jpg Kestrel Backtest Jobs, Parameter Sweeps and Honest Walk-Forward How do you run hundreds of backtests without cheating, without a network, and without fooling yourself? https://www.youtube.com/embed/Myr5kcD89y8 843 2026-10-31T09:37:53.651Z 4 yes backtesting walk-forward analysis parameter sweep C++20 algorithmic trading quant development job queue worker pool overfitting Sharpe ratio Kestrel NSE Indian markets systems programming out-of-sample testing https://easy-explain-app.vercel.app/lesson/tdfIdEJjsjA 2026-10-31 0.7 https://i.ytimg.com/vi/tdfIdEJjsjA/hqdefault.jpg https://i.ytimg.com/vi/tdfIdEJjsjA/hqdefault.jpg Backtest Metrics: Welford Sharpe, Sortino and Drawdown in C++ Four equity points can produce a Sharpe ratio above 12. Here is why, and how Kestrel's BacktestMetrics computes Sharpe, Sortino, volatility, CAGR and drawdown in one streaming pass. https://www.youtube.com/embed/tdfIdEJjsjA 796 2026-10-31T09:37:40.638Z 5 yes backtest metrics sharpe ratio sortino ratio max drawdown welford algorithm streaming variance C++20 quant developer algorithmic trading backtesting CAGR equity curve Kestrel Indian markets risk-adjusted return numerical stability https://easy-explain-app.vercel.app/lesson/48rKkAOY_Yw 2026-10-31 0.7 https://i.ytimg.com/vi/48rKkAOY_Yw/hqdefault.jpg https://i.ytimg.com/vi/48rKkAOY_Yw/hqdefault.jpg Building an IV Surface: Parity Forward, Black-76, 25-Delta Wings How do you turn a noisy option chain into one clean volatility number? In this episode of Inside Kestrel we build an implied volatility surface from scratch: a put-call parity forward, Black-76 inversion, ATM IV, 25-delta risk reversal and butterfly, and constant-maturity IV. https://www.youtube.com/embed/48rKkAOY_Yw 758 2026-10-31T09:37:27.386Z 12 yes implied volatility IV surface put-call parity forward price Black-76 25 delta risk reversal butterfly volatility smile total variance constant maturity volatility Nifty options quant developer C++20 options pricing calendar arbitrage Inside Kestrel https://easy-explain-app.vercel.app/lesson/YkKtVWDa2BY 2026-10-31 0.7 https://i.ytimg.com/vi/YkKtVWDa2BY/hqdefault.jpg https://i.ytimg.com/vi/YkKtVWDa2BY/hqdefault.jpg Black-Scholes, Greeks and Implied Volatility by Bisection in C++ Why does a production options library solve implied volatility with bisection instead of Newton's method? We read Kestrel's Black-Scholes code line by line. https://www.youtube.com/embed/YkKtVWDa2BY 755 2026-10-31T09:37:11.108Z 1 yes black scholes implied volatility bisection method newton method option greeks delta gamma vega theta C++20 quant developer options pricing put call parity Kestrel algorithmic trading platform NSE options numerical methods erfc normal cdf no arbitrage bounds https://easy-explain-app.vercel.app/lesson/l31ennq60ME 2026-10-31 0.7 https://i.ytimg.com/vi/l31ennq60ME/hqdefault.jpg https://i.ytimg.com/vi/l31ennq60ME/hqdefault.jpg The Indian Cost Model: Brokerage, STT, GST and Every Fee Per Fill A single option trade can look cheap and still pay seven different charges. In this episode of Inside Kestrel we build the Indian cost model line by line. https://www.youtube.com/embed/l31ennq60ME 792 2026-10-31T09:36:57.342Z 11 yes Indian cost model STT brokerage GST on brokerage algorithmic trading C++20 backtesting costs NSE transaction charges SEBI fee stamp duty partial fills NIFTY options NIFTY futures quant developer Kestrel https://easy-explain-app.vercel.app/lesson/nVIKJndn4Is 2026-10-31 0.7 https://i.ytimg.com/vi/nVIKJndn4Is/hqdefault.jpg https://i.ytimg.com/vi/nVIKJndn4Is/hqdefault.jpg The Simulated Venue: How Kestrel Fills Orders Without Lookahead When exactly does a simulated order fill, and at what price? A backtest that fills at the wrong bar quietly lies to you. https://www.youtube.com/embed/nVIKJndn4Is 740 2026-10-31T09:36:37.868Z 18 yes backtesting lookahead bias fill simulation simulated venue C++20 algorithmic trading order matching slippage limit order stop order participation rate quant developer NSE Indian markets event-driven engine Kestrel https://easy-explain-app.vercel.app/lesson/Y1bAzh2dNCA 2026-10-30 0.7 https://i.ytimg.com/vi/Y1bAzh2dNCA/hqdefault.jpg https://i.ytimg.com/vi/Y1bAzh2dNCA/hqdefault.jpg Positions and P&L in C++: Average Cost, Flips and Futures Variation Margin How does a trading engine know what you own, what you paid, and what you have made? Kestrel answers with one small function. https://www.youtube.com/embed/Y1bAzh2dNCA 705 2026-10-30T08:40:02.512Z 10 yes average cost accounting position manager variation margin futures accounting realized pnl unrealized pnl C++20 algorithmic trading backtesting engine equity identity Indian markets NSE futures quant developer trading system design Kestrel position flip https://easy-explain-app.vercel.app/lesson/K82HXlABkK4 2026-10-30 0.7 https://i.ytimg.com/vi/K82HXlABkK4/hqdefault.jpg https://i.ytimg.com/vi/K82HXlABkK4/hqdefault.jpg OMS and Risk: Normalise, Check, Then Send (Kestrel Code Walkthrough) What happens between an alpha saying "sell 60" and an order reaching a venue? In Kestrel, every order is normalised, risk-checked in a fixed order, and only then sent. https://www.youtube.com/embed/K82HXlABkK4 673 2026-10-30T08:39:46.177Z 11 yes C++20 algorithmic trading order management system OMS risk management kill switch Kestrel NSE Indian markets state machine fat finger price collar trading engine systems programming event-driven lot size tick size https://easy-explain-app.vercel.app/lesson/IWycUa9mA7U 2026-10-30 0.7 https://i.ytimg.com/vi/IWycUa9mA7U/hqdefault.jpg https://i.ytimg.com/vi/IWycUa9mA7U/hqdefault.jpg Kestrel Alpha Framework: Interfaces, Gates and Target Positions How does one C++ strategy object run unchanged in backtest, paper and live trading? Through a narrow interface, a chain of gates, and a habit of trading to targets instead of deltas. https://www.youtube.com/embed/IWycUa9mA7U 770 2026-10-30T08:39:30.404Z 3 yes C++20 algorithmic trading trading system design alpha framework target position order gating interface design fake context testing event-driven engine NSE Indian markets quant developer Kestrel backtesting engine idempotent orders software architecture https://easy-explain-app.vercel.app/lesson/uqtCDPxQnqg 2026-10-30 0.7 https://i.ytimg.com/vi/uqtCDPxQnqg/hqdefault.jpg https://i.ytimg.com/vi/uqtCDPxQnqg/hqdefault.jpg Inside Kestrel: The Per-Bar Pipeline, Derived Bars and Expiry One bar arrives. What happens next, and in what order? In this episode we walk through Kestrel's handleBar pipeline in a C++20 trading engine, and see why the order of steps prevents lookahead bugs. https://www.youtube.com/embed/uqtCDPxQnqg 727 2026-10-30T08:39:14.502Z 14 yes C++20 algorithmic trading trading engine event-driven architecture backtesting derived bars option expiry expiry settlement lookahead bias Kestrel NSE systems programming std::variant RAII paper trading https://easy-explain-app.vercel.app/lesson/ACqV2HMDSWw 2026-10-30 0.7 https://i.ytimg.com/vi/ACqV2HMDSWw/hqdefault.jpg https://i.ytimg.com/vi/ACqV2HMDSWw/hqdefault.jpg The Event Bus and K-Way Merge: Why Every Backtest Is Bit-Identical Run the same backtest twice and get the exact same bytes. How? https://www.youtube.com/embed/ACqV2HMDSWw 765 2026-10-30T08:38:58.303Z 15 yes C++20 event bus k-way merge determinism backtesting engine priority queue single thread ownership concurrency deadlock snapshot algorithmic trading Kestrel quant developer systems programming Indian markets reproducible backtest https://easy-explain-app.vercel.app/lesson/CeyDpWRj1Ls 2026-10-30 0.7 https://i.ytimg.com/vi/CeyDpWRj1Ls/hqdefault.jpg https://i.ytimg.com/vi/CeyDpWRj1Ls/hqdefault.jpg The History Store: Bar Files That Never Corrupt (Kestrel Ep. 4) What happens to your backtest when a data file is half written, or when every bar is one minute too early? In this episode we open Kestrel's history store and see how it survives crashes and concurrent writers. https://www.youtube.com/embed/CeyDpWRj1Ls 783 2026-10-30T08:38:43.311Z 9 yes C++20 Kestrel history store backtesting market data atomic file write fsync flock gzip CSV min-heap merge NSE FYERS systems programming data engineering crash safety Manim https://easy-explain-app.vercel.app/lesson/cPe2E0X15Gc 2026-10-30 0.7 https://i.ytimg.com/vi/cPe2E0X15Gc/hqdefault.jpg https://i.ytimg.com/vi/cPe2E0X15Gc/hqdefault.jpg Kestrel Core Vocabulary: Ticks, IST Time and the Run Config Before any engine logic, a trading system needs words it can trust. This episode walks through Kestrel's base types, the variant-index trick, tick rounding with an epsilon, hand-written IST time, and the three-file run config with its strict validator. https://www.youtube.com/embed/cPe2E0X15Gc 752 2026-10-30T08:38:27.465Z 3 yes C++20 algorithmic trading Kestrel NSE IST time tick size std::variant floating point config validation event-driven engine Indian markets FYERS quant developer systems programming JSON config https://easy-explain-app.vercel.app/lesson/j-wYUGibQjc 2026-10-30 0.7 https://i.ytimg.com/vi/j-wYUGibQjc/hqdefault.jpg https://i.ytimg.com/vi/j-wYUGibQjc/hqdefault.jpg Kestrel Build System: CMake Modules and a Zero-Comment Rule What happens when a C++20 trading codebase enforces its own boundaries? Kestrel turns every directory into a static library, makes dependency cycles impossible by design, and bans comments with a script. https://www.youtube.com/embed/j-wYUGibQjc 751 2026-10-30T08:38:13.204Z 8 yes CMake tutorial C++20 static libraries build system dependency graph acyclic dependencies AddressSanitizer UBSan compiler warnings Werror code style no comments rule C++ project structure algorithmic trading platform Kestrel systems programming software architecture static initialization https://easy-explain-app.vercel.app/lesson/EgmlrRIfqY8 2026-10-30 0.7 https://i.ytimg.com/vi/EgmlrRIfqY8/hqdefault.jpg https://i.ytimg.com/vi/EgmlrRIfqY8/hqdefault.jpg Kestrel in One Picture: One Engine, Three Modes What if your backtest and your live trading ran the exact same code? Kestrel, a C++20 trading platform for Indian markets, does exactly that. https://www.youtube.com/embed/EgmlrRIfqY8 756 2026-10-30T08:37:55.905Z 23 yes C++20 algorithmic trading backtesting paper trading live trading software architecture event-driven NSE BSE MCX FYERS API CMake trading engine systems programming Kestrel quant developer https://easy-explain-app.vercel.app/lesson/5N5y5aq_faE 2026-10-30 0.7 https://i.ytimg.com/vi/5N5y5aq_faE/hqdefault.jpg https://i.ytimg.com/vi/5N5y5aq_faE/hqdefault.jpg Prompt Injection in 50 Seconds: How AI Agents Get Hijacked A web page can secretly give your AI agent orders. Here is prompt injection in 50 seconds, plus two defenses that actually help. https://www.youtube.com/embed/5N5y5aq_faE 53 2026-10-30T08:37:42.256Z 39 yes prompt injection indirect prompt injection LLM security AI agents AI safety guardrails privilege separation human in the loop generative AI LLM engineering tool calling agent security shorts machine learning Manim https://easy-explain-app.vercel.app/lesson/GRGM2xSkTN0 2026-10-29 0.7 https://i.ytimg.com/vi/GRGM2xSkTN0/hqdefault.jpg https://i.ytimg.com/vi/GRGM2xSkTN0/hqdefault.jpg Why 95% Accurate AI Agents Still Fail Most Tasks Your agent is right 95 percent of the time per step, so why does it fail so often? Because the errors multiply. https://www.youtube.com/embed/GRGM2xSkTN0 54 2026-10-29T09:49:19.536Z 13 yes AI agents LLM agents agent reliability compounding errors 0.95 to the 10 ReAct generative AI LLM engineering tool calling agent verification AI explained machine learning Manim prompt engineering shorts https://easy-explain-app.vercel.app/lesson/3R6ydAxWMn4 2026-10-29 0.7 https://i.ytimg.com/vi/3R6ydAxWMn4/hqdefault.jpg https://i.ytimg.com/vi/3R6ydAxWMn4/hqdefault.jpg DPO vs RLHF: What Got Removed From Alignment RLHF needs a reward model, PPO, and a lot of moving parts. DPO deletes most of that and trains with one simple loss. https://www.youtube.com/embed/3R6ydAxWMn4 69 2026-10-29T09:49:14.533Z 130 yes DPO RLHF direct preference optimization PPO reward model LLM alignment fine-tuning preference learning generative AI LLM engineering machine learning AI explained Bradley-Terry language models Shorts https://easy-explain-app.vercel.app/lesson/LBuQcNXnyy8 2026-10-29 0.7 https://i.ytimg.com/vi/LBuQcNXnyy8/hqdefault.jpg https://i.ytimg.com/vi/LBuQcNXnyy8/hqdefault.jpg LoRA in One Minute: 16.8 Million Parameters vs 65,536 One weight matrix, 16.8 million numbers. LoRA trains just 65,536 of them. https://www.youtube.com/embed/LBuQcNXnyy8 57 2026-10-29T09:49:09.969Z 228 yes LoRA low rank adaptation fine-tuning LLM PEFT parameter efficient fine-tuning low-rank matrices machine learning generative AI LLM engineering Manim deep learning transformers AI explained Shorts https://easy-explain-app.vercel.app/lesson/Fv6XxKEjy80 2026-10-29 0.7 https://i.ytimg.com/vi/Fv6XxKEjy80/hqdefault.jpg https://i.ytimg.com/vi/Fv6XxKEjy80/hqdefault.jpg Why Keyword Search Still Beats Embeddings Sometimes Embeddings are powerful, but they can miss an exact error code that old-school BM25 finds instantly. https://www.youtube.com/embed/Fv6XxKEjy80 54 2026-10-29T09:49:05.093Z 20 yes keyword search BM25 embeddings vector search hybrid search reciprocal rank fusion RRF RAG LLM engineering generative AI semantic search dense retrieval information retrieval error code search Shorts https://easy-explain-app.vercel.app/lesson/7uW_i14xvaY 2026-10-29 0.7 https://i.ytimg.com/vi/7uW_i14xvaY/hqdefault.jpg https://i.ytimg.com/vi/7uW_i14xvaY/hqdefault.jpg Product Quantization: Shrink a 3072-Byte Vector to 96 Bytes One embedding takes 3072 bytes. Product quantization stores it in 96, which is 32 times smaller. https://www.youtube.com/embed/7uW_i14xvaY 65 2026-10-29T09:49:00.008Z 342 yes product quantization vector search embeddings vector database ANN approximate nearest neighbor k-means codebook compression recall RAG LLM engineering generative AI faiss machine learning shorts https://easy-explain-app.vercel.app/lesson/ibUk_bQsH4Q 2026-10-29 0.7 https://i.ytimg.com/vi/ibUk_bQsH4Q/hqdefault.jpg https://i.ytimg.com/vi/ibUk_bQsH4Q/hqdefault.jpg Cosine vs Dot Product: The Search Ranking Trap Your vector search might be ranking the wrong result, and vector length is why. https://www.youtube.com/embed/ibUk_bQsH4Q 59 2026-10-29T09:48:55.273Z 160 yes cosine similarity dot product embeddings vector search normalize embeddings RAG LLM engineering vector database semantic search machine learning generative AI similarity metrics nearest neighbor shorts manim https://easy-explain-app.vercel.app/lesson/2dy7e5K9mMA 2026-10-29 0.7 https://i.ytimg.com/vi/2dy7e5K9mMA/hqdefault.jpg https://i.ytimg.com/vi/2dy7e5K9mMA/hqdefault.jpg Why Chatbots Confidently Make Things Up Your chatbot just cited a paper that doesn't exist. Here's why. https://www.youtube.com/embed/2dy7e5K9mMA 47 2026-10-29T09:48:50.580Z 37 yes LLM hallucination why chatbots lie AI makes things up fabricated citations next token prediction RAG retrieval augmented generation grounding generative AI course LLM engineering ChatGPT hallucination transformers AI explained machine learning Shorts https://easy-explain-app.vercel.app/lesson/RggtSEH4g9g 2026-10-29 0.7 https://i.ytimg.com/vi/RggtSEH4g9g/hqdefault.jpg https://i.ytimg.com/vi/RggtSEH4g9g/hqdefault.jpg Why AI Types Slowly: The KV Cache Explained Why does an LLM generate text one word at a time, and why does it get slower as the conversation grows? The answer is the KV cache. https://www.youtube.com/embed/RggtSEH4g9g 54 2026-10-29T09:48:45.735Z 140 yes KV cache LLM inference why LLMs are slow decoding memory bandwidth transformer attention GPT generative AI LLM engineering token generation GPU memory machine learning AI explained Shorts https://easy-explain-app.vercel.app/lesson/XpNpP99kQ2I 2026-10-29 0.7 https://i.ytimg.com/vi/XpNpP99kQ2I/hqdefault.jpg https://i.ytimg.com/vi/XpNpP99kQ2I/hqdefault.jpg Temperature in 50 Seconds: How LLMs Choose the Next Word Why does an AI answer differently every time? Temperature is the dial behind it. https://www.youtube.com/embed/XpNpP99kQ2I 59 2026-10-29T09:48:40.406Z 119 yes temperature LLM temperature softmax logits sampling next token prediction generative AI LLM engineering top-p top-k transformers machine learning AI explained Manim shorts https://easy-explain-app.vercel.app/lesson/unxHpAvwR0c 2026-10-29 0.7 https://i.ytimg.com/vi/unxHpAvwR0c/hqdefault.jpg https://i.ytimg.com/vi/unxHpAvwR0c/hqdefault.jpg Capstone 4: Deploy a Tool-Using AI Agent to Production (FastAPI, Docker, Guardrails) Build and ship a production-style tool-using agent: tool calling, memory, step and cost budgets, guardrails, human approval, FastAPI streaming, Docker, tracing, CI evals, canary and rollback. https://www.youtube.com/embed/unxHpAvwR0c 770 2026-10-29T09:48:35.806Z 12 yes ai agent tool calling llm agents fastapi docker deploy llm mcp vllm prompt injection guardrails agent budgets human in the loop llm evals ci canary deployment generative ai course llm engineering https://easy-explain-app.vercel.app/lesson/zmteR_v5BAo 2026-10-28 0.7 https://i.ytimg.com/vi/zmteR_v5BAo/hqdefault.jpg https://i.ytimg.com/vi/zmteR_v5BAo/hqdefault.jpg Capstone: Fine-Tune a Small Open Model End to End (QLoRA + TRL) Fine-tune a small open model for a narrow task, the right way: baseline first, clean data, QLoRA with TRL, honest evaluation, merge, quantize, and ship with a model card. https://www.youtube.com/embed/zmteR_v5BAo 752 2026-10-28T09:05:41.475Z 9 yes fine-tuning LLM QLoRA TRL SFT LoRA DPO Hugging Face structured extraction LLM evaluation model card merge adapters quantization data leakage capstone project generative AI course LLM engineering https://easy-explain-app.vercel.app/lesson/6T6zVctD2EM 2026-10-28 0.7 https://i.ytimg.com/vi/6T6zVctD2EM/hqdefault.jpg https://i.ytimg.com/vi/6T6zVctD2EM/hqdefault.jpg Capstone 2: Build a Repo-Aware Coding Assistant (Python, Sandboxed, Tested) What does it take to build a coding assistant that edits your code and proves it works? In this capstone we build one from scratch in plain Python. https://www.youtube.com/embed/6T6zVctD2EM 785 2026-10-28T09:05:25.083Z 37 yes coding assistant build coding assistant LLM agent ReAct loop AST chunking tree-sitter hybrid search unified diff sandbox prompt injection generative AI course LLM engineering Python agent tool calling code RAG AI capstone project https://easy-explain-app.vercel.app/lesson/8cwn3yn-ooc 2026-10-28 0.7 https://i.ytimg.com/vi/8cwn3yn-ooc/hqdefault.jpg https://i.ytimg.com/vi/8cwn3yn-ooc/hqdefault.jpg Build a Document Q&A RAG Bot: Hybrid Search, Reranking, Citations, Evals Most RAG demos work on three hand-picked questions and fall apart on the fourth. In this capstone we build a document Q&A bot that you can actually trust, and we prove it with an eval set. https://www.youtube.com/embed/8cwn3yn-ooc 854 2026-10-28T09:05:09.409Z 5 yes RAG tutorial document Q&A bot build a RAG app hybrid search reciprocal rank fusion cross-encoder reranking RAG evaluation recall at k faithfulness FastAPI Gradio vector database LLM engineering chunking strategies citations and abstention generative AI course https://easy-explain-app.vercel.app/lesson/i8pR2yRAT4A 2026-10-28 0.7 https://i.ytimg.com/vi/i8pR2yRAT4A/hqdefault.jpg https://i.ytimg.com/vi/i8pR2yRAT4A/hqdefault.jpg LLM Guardrails and Prompt Injection: Securing and Monitoring AI in Production Your LLM app passed every eval, so why can one pasted web page make it leak data? https://www.youtube.com/embed/i8pR2yRAT4A 844 2026-10-28T09:04:54.626Z 22 yes prompt injection LLM guardrails LLM security AI safety indirect prompt injection LLM monitoring LLM observability excessive agency PII redaction jailbreaks defense in depth canary release LLM as judge production LLM generative AI course AI agents security https://easy-explain-app.vercel.app/lesson/CDgof5l1jDA 2026-10-28 0.7 https://i.ytimg.com/vi/CDgof5l1jDA/hqdefault.jpg https://i.ytimg.com/vi/CDgof5l1jDA/hqdefault.jpg Serving LLMs with vLLM: Continuous Batching and PagedAttention Explained Why does the same GPU serve several times more users with vLLM? Two ideas: continuous batching and PagedAttention. https://www.youtube.com/embed/CDgof5l1jDA 877 2026-10-28T09:04:37.590Z 2 yes vLLM PagedAttention continuous batching LLM serving LLM inference KV cache GPU memory speculative decoding tensor parallelism chunked prefill prefix caching OpenAI compatible server LLM engineering generative AI course inference optimization throughput latency https://easy-explain-app.vercel.app/lesson/7b1C-BeC5wU 2026-10-28 0.7 https://i.ytimg.com/vi/7b1C-BeC5wU/hqdefault.jpg https://i.ytimg.com/vi/7b1C-BeC5wU/hqdefault.jpg Quantization Explained: INT8, INT4, GPTQ and AWQ for LLMs How do you fit a 70 billion parameter model into a fraction of the memory without wrecking it? Quantization. https://www.youtube.com/embed/7b1C-BeC5wU 913 2026-10-28T09:04:20.259Z 20 yes quantization INT8 INT4 GPTQ AWQ LLM quantization GGUF llama.cpp post-training quantization FP8 bitsandbytes Hugging Face LLM inference model compression generative AI course memory bandwidth https://easy-explain-app.vercel.app/lesson/NhA_XtqraH0 2026-10-28 0.7 https://i.ytimg.com/vi/NhA_XtqraH0/hqdefault.jpg https://i.ytimg.com/vi/NhA_XtqraH0/hqdefault.jpg Latency and Cost Math for LLM Applications: Predict Your Bill Your LLM demo costs a fraction of a cent. Your production bill says otherwise. Learn the math that explains why. https://www.youtube.com/embed/NhA_XtqraH0 831 2026-10-28T09:04:02.291Z 0 yes LLM latency LLM cost time to first token TTFT tokens per second prompt caching LLM pricing RAG cost AI agent cost p95 latency LLM engineering generative AI course streaming LLM model routing token pricing LLM optimization https://easy-explain-app.vercel.app/lesson/TqUMJN2kYFI 2026-10-28 0.7 https://i.ytimg.com/vi/TqUMJN2kYFI/hqdefault.jpg https://i.ytimg.com/vi/TqUMJN2kYFI/hqdefault.jpg Model Context Protocol (MCP) Explained: One Standard for AI Tools Every AI app needs tools, and every tool needs every app. MCP is the open protocol that turns that mess into one clean standard. https://www.youtube.com/embed/TqUMJN2kYFI 804 2026-10-28T09:03:48.017Z 125 yes MCP Model Context Protocol LLM tools AI agents function calling JSON-RPC MCP server Python MCP SDK tool poisoning prompt injection generative AI course LLM engineering stdio transport streamable HTTP agent security https://easy-explain-app.vercel.app/lesson/oBs2xT2_gJU 2026-10-28 0.7 https://i.ytimg.com/vi/oBs2xT2_gJU/hqdefault.jpg https://i.ytimg.com/vi/oBs2xT2_gJU/hqdefault.jpg Multi-Agent Systems: When They Help and When They Hurt More agents is not automatically better. Learn when splitting work across agents pays off, and when one agent wins. https://www.youtube.com/embed/oBs2xT2_gJU 775 2026-10-28T09:03:34.282Z 39 yes multi-agent systems LLM agents orchestrator worker agent orchestration AI agents tutorial supervisor agent agent handoffs context isolation generative AI course LLM engineering agent cost analysis python agents critic agent parallel agents single agent vs multi agent https://easy-explain-app.vercel.app/lesson/bz2vS3a1_TM 2026-10-28 0.7 https://i.ytimg.com/vi/bz2vS3a1_TM/hqdefault.jpg https://i.ytimg.com/vi/bz2vS3a1_TM/hqdefault.jpg Agent Memory Explained: Short-Term, Long-Term and Context Management Your agent forgets everything the moment the context window closes. Here is how to fix that. https://www.youtube.com/embed/bz2vS3a1_TM 789 2026-10-28T09:03:19.903Z 4 yes agent memory LLM agents context engineering short-term memory long-term memory vector store memory conversation summarization context window prompt caching RAG vs memory AI agents tutorial generative AI course LLM engineering sliding window memory episodic memory compaction https://easy-explain-app.vercel.app/lesson/IJum0tlGedQ 2026-10-27 0.7 https://i.ytimg.com/vi/IJum0tlGedQ/hqdefault.jpg https://i.ytimg.com/vi/IJum0tlGedQ/hqdefault.jpg The ReAct Loop, Planning and Reflection: How AI Agents Actually Think Build an AI agent from scratch: the ReAct loop explained, plus planning, reflection and the failure modes that sink real agents. https://www.youtube.com/embed/IJum0tlGedQ 798 2026-10-27T10:15:07.536Z 367 yes ReAct agent ReAct loop AI agents LLM agents agent planning reflection Reflexion self-critique tool calling agent from scratch Python agent tutorial LLM engineering generative AI course plan and execute agent failure modes max steps agent agent trace logging https://easy-explain-app.vercel.app/lesson/5BBBJGuUpmY 2026-10-27 0.7 https://i.ytimg.com/vi/5BBBJGuUpmY/hqdefault.jpg https://i.ytimg.com/vi/5BBBJGuUpmY/hqdefault.jpg Tool and Function Calling: How LLMs Actually Use Tools A language model can only write text, so how does it check the weather, do exact math, or call your API? The answer is tool calling: the model writes a request, and your code acts. https://www.youtube.com/embed/5BBBJGuUpmY 791 2026-10-27T10:14:53.778Z 8 yes tool calling function calling LLM tools AI agents JSON schema Pydantic parallel tool calls constrained decoding prompt injection LLM engineering generative AI course python tool loop structured output agent tools tool use LLM https://easy-explain-app.vercel.app/lesson/3X79mWN6jH4 2026-10-27 0.7 https://i.ytimg.com/vi/3X79mWN6jH4/hqdefault.jpg https://i.ytimg.com/vi/3X79mWN6jH4/hqdefault.jpg Fine-Tune vs RAG vs Prompt: How to Decide (A Practical Framework) Should you prompt, retrieve, or fine-tune? Most teams pick the wrong tool first. This lesson gives you a clear decision ladder, a cost comparison with hypothetical prices, four case studies, and Python code to decide with evidence. https://www.youtube.com/embed/3X79mWN6jH4 771 2026-10-27T10:14:39.127Z 8 yes fine-tuning vs RAG RAG vs fine-tuning prompt engineering LLM engineering when to fine-tune retrieval augmented generation LoRA generative AI course LLM cost comparison model distillation LLM evaluation structured output AI tutorial decision framework few-shot prompting https://easy-explain-app.vercel.app/lesson/FrgRkrSXLfE 2026-10-27 0.7 https://i.ytimg.com/vi/FrgRkrSXLfE/hqdefault.jpg https://i.ytimg.com/vi/FrgRkrSXLfE/hqdefault.jpg DPO Derived: Preference Optimization Without RL What if you could get the benefits of RLHF with a single, simple classification-style loss? In this lesson we derive Direct Preference Optimization from scratch. https://www.youtube.com/embed/FrgRkrSXLfE 859 2026-10-27T10:14:25.585Z 7 yes DPO direct preference optimization RLHF PPO preference optimization LLM alignment Bradley-Terry KL divergence TRL DPOTrainer reward model fine-tuning LLM IPO KTO ORPO generative AI course LLM engineering PyTorch https://easy-explain-app.vercel.app/lesson/1TtwvD2JivM 2026-10-27 0.7 https://i.ytimg.com/vi/1TtwvD2JivM/hqdefault.jpg https://i.ytimg.com/vi/1TtwvD2JivM/hqdefault.jpg RLHF Explained: Reward Models, KL Penalty and PPO Step by Step How does a raw language model become a helpful assistant? Reinforcement learning from human feedback, explained with real numbers. https://www.youtube.com/embed/1TtwvD2JivM 902 2026-10-27T10:14:11.584Z 16 yes RLHF reinforcement learning from human feedback reward model PPO proximal policy optimization Bradley-Terry KL divergence reward hacking sycophancy LLM alignment instruction tuning GRPO RLVR DPO PyTorch fine-tuning LLMs InstructGPT https://easy-explain-app.vercel.app/lesson/ZfeWoghxgII 2026-10-27 0.7 https://i.ytimg.com/vi/ZfeWoghxgII/hqdefault.jpg https://i.ytimg.com/vi/ZfeWoghxgII/hqdefault.jpg QLoRA Explained: Fine-Tune a 7B Model on One GPU with 4-Bit Quantization A 7 billion parameter model needs over 100 gigabytes to fine-tune the normal way. QLoRA does it in a fraction of that. Here is exactly how. https://www.youtube.com/embed/ZfeWoghxgII 775 2026-10-27T10:13:57.686Z 44 yes QLoRA quantized training LoRA NF4 4-bit quantization bitsandbytes PEFT TRL fine-tune LLM double quantization paged optimizers gradient checkpointing bf16 mixed precision LLM fine-tuning memory Adam optimizer memory generative AI course https://easy-explain-app.vercel.app/lesson/0FyEOCwDgts 2026-10-27 0.7 https://i.ytimg.com/vi/0FyEOCwDgts/hqdefault.jpg https://i.ytimg.com/vi/0FyEOCwDgts/hqdefault.jpg LoRA Math Explained: Low-Rank Adaptation Derived From Scratch Fine-tune a huge model by training less than half a percent of one weight matrix? Here is exactly why LoRA works. https://www.youtube.com/embed/0FyEOCwDgts 890 2026-10-27T10:13:44.539Z 15 yes LoRA low-rank adaptation LoRA explained LoRA math PEFT parameter efficient fine-tuning fine-tuning LLM PyTorch LoRA Hugging Face PEFT LoRA from scratch rank and alpha merge LoRA adapter generative AI course LLM engineering matrix factorization https://easy-explain-app.vercel.app/lesson/u8pitUokrfE 2026-10-27 0.7 https://i.ytimg.com/vi/u8pitUokrfE/hqdefault.jpg https://i.ytimg.com/vi/u8pitUokrfE/hqdefault.jpg Supervised Fine-Tuning Explained: Chat Templates, Loss Masking and Code How does a model that only predicts the next token become an assistant that follows instructions? In this lesson we build supervised fine-tuning from the ground up. https://www.youtube.com/embed/u8pitUokrfE 707 2026-10-27T10:13:30.136Z 2 yes supervised fine-tuning SFT instruction tuning chat template loss masking TRL SFTTrainer Hugging Face LLM fine-tuning catastrophic forgetting cross-entropy ignore_index -100 synthetic data instruction dataset generative AI course LLM engineering https://easy-explain-app.vercel.app/lesson/uActEaJAYkk 2026-10-27 0.7 https://i.ytimg.com/vi/uActEaJAYkk/hqdefault.jpg https://i.ytimg.com/vi/uActEaJAYkk/hqdefault.jpg Evaluating RAG: Metrics, Test Sets and 8 Failure Modes Your RAG demo looks great, so why does it fail in production? Because you are probably judging it by eyeballing a few answers. https://www.youtube.com/embed/uActEaJAYkk 746 2026-10-27T10:13:16.951Z 17 yes RAG evaluation evaluating RAG RAG failure modes recall at k MRR nDCG faithfulness LLM as a judge retrieval augmented generation RAG test set context precision context recall answer relevance lost in the middle LLM engineering generative AI course https://easy-explain-app.vercel.app/lesson/D5GscpT1ars 2026-10-27 0.7 https://i.ytimg.com/vi/D5GscpT1ars/hqdefault.jpg https://i.ytimg.com/vi/D5GscpT1ars/hqdefault.jpg Hybrid Search Explained: BM25 + Dense Vectors with Reciprocal Rank Fusion Your vector search just missed an exact error code. Here is why, and how hybrid search fixes it. https://www.youtube.com/embed/D5GscpT1ars 790 2026-10-27T10:13:03.321Z 38 yes hybrid search BM25 reciprocal rank fusion RRF dense retrieval RAG vector search rank_bm25 SPLADE sparse retrieval embeddings information retrieval LLM engineering semantic search keyword search retrieval augmented generation https://easy-explain-app.vercel.app/lesson/jjll_YlwJn0 2026-10-26 0.7 https://i.ytimg.com/vi/jjll_YlwJn0/hqdefault.jpg https://i.ytimg.com/vi/jjll_YlwJn0/hqdefault.jpg Reranking with Cross-Encoders: The Cheap Trick That Fixes RAG Search Your vector search finds roughly the right documents, but the best one is stuck at rank 7. Cross-encoder reranking fixes that. https://www.youtube.com/embed/jjll_YlwJn0 826 2026-10-26T08:35:46.204Z 5 yes reranking cross-encoder bi-encoder RAG retrieval augmented generation sentence-transformers ColBERT late interaction MRR recall at k two-stage retrieval vector search LLM engineering semantic search generative AI course LLM reranker https://easy-explain-app.vercel.app/lesson/T2bEWLvGz7M 2026-10-26 0.7 https://i.ytimg.com/vi/T2bEWLvGz7M/hqdefault.jpg https://i.ytimg.com/vi/T2bEWLvGz7M/hqdefault.jpg Retrieval and Query Rewriting: Fix RAG Before It Fails Your RAG bot has the right document and still answers wrong. The problem is often the query, not the index. https://www.youtube.com/embed/T2bEWLvGz7M 879 2026-10-26T08:35:32.944Z 22 yes RAG retrieval augmented generation query rewriting HyDE hypothetical document embeddings MMR maximal marginal relevance multi-query retrieval query decomposition query routing step-back prompting top-k retrieval metadata filtering LLM engineering vector search generative AI course https://easy-explain-app.vercel.app/lesson/N6x7Wnjq9SQ 2026-10-26 0.7 https://i.ytimg.com/vi/N6x7Wnjq9SQ/hqdefault.jpg https://i.ytimg.com/vi/N6x7Wnjq9SQ/hqdefault.jpg Chunking Strategies for RAG: How to Split Documents Right Your RAG system can only find what you chunked well. Learn fixed-size, recursive, structure-aware, semantic and parent-document chunking, with real numbers and Python. https://www.youtube.com/embed/N6x7Wnjq9SQ 821 2026-10-26T08:35:18.812Z 155 yes RAG chunking chunking strategies retrieval augmented generation recursive character splitting semantic chunking parent document retriever small to big retrieval LLM engineering vector search embeddings text splitting contextual retrieval document parsing generative AI course https://easy-explain-app.vercel.app/lesson/jnQoARknUXg 2026-10-26 0.7 https://i.ytimg.com/vi/jnQoARknUXg/hqdefault.jpg https://i.ytimg.com/vi/jnQoARknUXg/hqdefault.jpg RAG Explained: How to Ground LLMs in Your Own Data What if your LLM could look up the answer instead of guessing it? https://www.youtube.com/embed/jnQoARknUXg 747 2026-10-26T08:35:05.063Z 79 yes RAG retrieval augmented generation LLM grounding LLMs vector search embeddings hallucination knowledge cutoff prompt template citations python RAG RAG from scratch AI engineering generative AI course semantic search vector database chatbot with documents https://easy-explain-app.vercel.app/lesson/ZtTkvM-SGBM 2026-10-26 0.7 https://i.ytimg.com/vi/ZtTkvM-SGBM/hqdefault.jpg https://i.ytimg.com/vi/ZtTkvM-SGBM/hqdefault.jpg Vector Databases in Practice: Filtering, Sizing, and Real Code An ANN index is not a database. Learn what a vector database really adds, and how to use one well. https://www.youtube.com/embed/ZtTkvM-SGBM 794 2026-10-26T08:34:50.735Z 6 yes vector database vector search embeddings pgvector Chroma LanceDB FAISS metadata filtering HNSW semantic search RAG LLM engineering generative AI multi-tenancy vector database tutorial approximate nearest neighbor https://easy-explain-app.vercel.app/lesson/744E2X-idks 2026-10-26 0.7 https://i.ytimg.com/vi/744E2X-idks/hqdefault.jpg https://i.ytimg.com/vi/744E2X-idks/hqdefault.jpg ANN Indexes Explained: HNSW, IVF and Product Quantization How do vector databases search millions of embeddings in milliseconds? Learn how HNSW graphs, IVF clustering and product quantization trade a little recall for huge speed and memory savings. https://www.youtube.com/embed/744E2X-idks 929 2026-10-26T08:34:35.072Z 37 yes ANN index HNSW IVF product quantization FAISS vector search approximate nearest neighbor vector database recall@10 nprobe efSearch IVFPQ embeddings RAG LLM engineering hnswlib https://easy-explain-app.vercel.app/lesson/m3l8mcvY2d4 2026-10-26 0.7 https://i.ytimg.com/vi/m3l8mcvY2d4/hqdefault.jpg https://i.ytimg.com/vi/m3l8mcvY2d4/hqdefault.jpg Cosine vs Dot Product: How Exact Vector Search Really Works Your vector search can return the wrong results even when the embedding model is perfect. The cause is often a mismatch between dot product, cosine similarity and Euclidean distance. https://www.youtube.com/embed/m3l8mcvY2d4 783 2026-10-26T08:34:20.241Z 43 yes cosine similarity dot product euclidean distance nearest neighbor search k-NN vector search embeddings brute force search curse of dimensionality recall at k numpy PyTorch RAG LLM engineering normalization FAISS https://easy-explain-app.vercel.app/lesson/EYVVDHaVXxY 2026-10-26 0.7 https://i.ytimg.com/vi/EYVVDHaVXxY/hqdefault.jpg https://i.ytimg.com/vi/EYVVDHaVXxY/hqdefault.jpg Sentence Embeddings and Contrastive Training: How Meaning Becomes Math How does a computer know that "How do I reset my password?" and "I forgot my login" mean the same thing? Sentence embeddings turn text into points in space, and contrastive training teaches them where to go. https://www.youtube.com/embed/EYVVDHaVXxY 767 2026-10-26T08:34:07.622Z 16 yes sentence embeddings contrastive learning InfoNCE in-batch negatives sentence-transformers text embeddings semantic search pooling mean pooling Matryoshka embeddings hard negatives temperature RAG LLM engineering PyTorch vector search https://easy-explain-app.vercel.app/lesson/sJRSjP-zGWk 2026-10-26 0.7 https://i.ytimg.com/vi/sJRSjP-zGWk/hqdefault.jpg https://i.ytimg.com/vi/sJRSjP-zGWk/hqdefault.jpg Structured Output and Evaluating Prompts Like Real Software Your LLM works in the demo, then breaks your parser in production. Here is how to fix both problems. https://www.youtube.com/embed/sJRSjP-zGWk 914 2026-10-26T08:33:52.664Z 4 yes structured output json mode llm structured output pydantic schema constrained decoding constrained decoding prompt evaluation llm evals llm as judge golden dataset prompt engineering regression testing prompts a/b testing prompts generative ai course llm engineering function calling json https://easy-explain-app.vercel.app/lesson/YY8UJ1Y8vJ8 2026-10-26 0.7 https://i.ytimg.com/vi/YY8UJ1Y8vJ8/hqdefault.jpg https://i.ytimg.com/vi/YY8UJ1Y8vJ8/hqdefault.jpg Chain-of-Thought and Reasoning Models: How LLMs Learn to Think Why does asking a model to think step by step change its answers? This lesson explains chain-of-thought prompting, self-consistency voting with a worked probability example, tree search ideas, and how reasoning models use reinforcement learning and test-time compute. https://www.youtube.com/embed/YY8UJ1Y8vJ8 768 2026-10-26T08:33:38.783Z 27 yes chain of thought reasoning models LLM reasoning self-consistency test-time compute prompt engineering think step by step reinforcement learning verifiable rewards thinking budget generative AI course LLM engineering majority vote LLM tree of thoughts chain of thought faithfulness AI explained https://easy-explain-app.vercel.app/lesson/IqdOibMocJc 2026-10-25 0.7 https://i.ytimg.com/vi/IqdOibMocJc/hqdefault.jpg https://i.ytimg.com/vi/IqdOibMocJc/hqdefault.jpg Few-Shot Prompting and In-Context Learning Explained (With Code) What if you could teach a language model a brand new task without changing a single weight? That is in-context learning, and it is the engine behind few-shot prompting. https://www.youtube.com/embed/IqdOibMocJc 739 2026-10-25T05:15:01.452Z 19 yes few-shot prompting in-context learning zero-shot prompting one-shot prompting prompt engineering LLM large language models kNN example selection dynamic few-shot TF-IDF prompt caching recency bias label bias generative AI course LLM engineering text classification with LLM https://easy-explain-app.vercel.app/lesson/S_9qB8jhqGk 2026-10-25 0.7 https://i.ytimg.com/vi/S_9qB8jhqGk/hqdefault.jpg https://i.ytimg.com/vi/S_9qB8jhqGk/hqdefault.jpg Prompt Patterns and System Prompts: Programming in Plain English Prompting is programming with natural language specs. Learn the patterns that make LLM output reliable. https://www.youtube.com/embed/S_9qB8jhqGk 769 2026-10-25T05:14:45.664Z 2 yes prompt engineering system prompt LLM prompt patterns chat messages instruction hierarchy prompt templates Jinja Python LLM prompt injection generative AI course LLM engineering structured prompts support email classifier delimiters https://easy-explain-app.vercel.app/lesson/zalzq8lOAKk 2026-10-25 0.7 https://i.ytimg.com/vi/zalzq8lOAKk/hqdefault.jpg https://i.ytimg.com/vi/zalzq8lOAKk/hqdefault.jpg Why LLMs Hallucinate (and What Actually Helps) Why does a language model confidently invent facts, names and citations? Because it is trained to produce plausible text, not true text. https://www.youtube.com/embed/zalzq8lOAKk 702 2026-10-25T05:14:32.189Z 7 yes LLM hallucination why LLMs hallucinate AI hallucinations explained self-consistency large language models generative AI course reduce hallucination RAG calibration temperature sampling fabricated citations LLM engineering ChatGPT hallucination machine learning grounding https://easy-explain-app.vercel.app/lesson/uJLSDw2xJ4w 2026-10-25 0.7 https://i.ytimg.com/vi/uJLSDw2xJ4w/hqdefault.jpg https://i.ytimg.com/vi/uJLSDw2xJ4w/hqdefault.jpg The KV Cache Explained: Why LLM Decoding Is Memory-Bound Why does a GPU that can do hundreds of trillions of operations per second still crawl through text one token at a time? The answer is the KV cache and memory bandwidth. https://www.youtube.com/embed/uJLSDw2xJ4w 813 2026-10-25T05:14:18.975Z 145 yes KV cache LLM inference memory-bound decoding prefill and decode roofline model arithmetic intensity grouped-query attention multi-query attention transformer PyTorch Hugging Face GPU memory bandwidth LLM engineering sliding window attention cache quantization https://easy-explain-app.vercel.app/lesson/oPNrMkhFJGg 2026-10-25 0.7 https://i.ytimg.com/vi/oPNrMkhFJGg/hqdefault.jpg https://i.ytimg.com/vi/oPNrMkhFJGg/hqdefault.jpg Context Windows and RoPE Explained: How LLMs Know Word Order How does a Transformer know which word came first? We derive Rotary Position Embeddings (RoPE) from scratch, work a 2D example with real numbers, and verify the relative-position property in PyTorch. https://www.youtube.com/embed/oPNrMkhFJGg 838 2026-10-25T05:14:02.970Z 49 yes RoPE rotary position embedding positional encoding context window long context transformer LLM attention position interpolation YaRN ALiBi lost in the middle PyTorch generative AI LLM engineering https://easy-explain-app.vercel.app/lesson/r3LnXDOJZ_8 2026-10-25 0.7 https://i.ytimg.com/vi/r3LnXDOJZ_8/hqdefault.jpg https://i.ytimg.com/vi/r3LnXDOJZ_8/hqdefault.jpg Sampling Explained: Greedy, Temperature, Top-k, Top-p (With Numbers) Why does the same LLM sound robotic in one app and wildly creative in another? The answer is not in the model. It is in how we pick the next token. https://www.youtube.com/embed/r3LnXDOJZ_8 783 2026-10-25T05:13:48.891Z 66 yes llm sampling temperature softmax top-p sampling nucleus sampling top-k sampling greedy decoding min-p sampling beam search repetition penalty logits softmax temperature pytorch sampling hugging face generate generative ai course llm engineering llm determinism https://easy-explain-app.vercel.app/lesson/vItWq_4QhSE 2026-10-25 0.7 https://i.ytimg.com/vi/vItWq_4QhSE/hqdefault.jpg https://i.ytimg.com/vi/vItWq_4QhSE/hqdefault.jpg Scaling Laws Explained: Why Bigger Models Keep Getting Better Why does making a language model bigger keep making it better, and so predictably? In this lesson we unpack scaling laws: power laws, the compute rule C ≈ 6ND, a worked GPU-days estimate for a 7B model, Chinchilla vs Kaplan, and why modern small models are over-trained. https://www.youtube.com/embed/vItWq_4QhSE 754 2026-10-25T05:13:30.776Z 7 yes scaling laws LLM scaling laws Chinchilla Kaplan scaling laws compute optimal training power law C=6ND large language models GPT training tokens per parameter test-time compute emergent abilities data wall generative AI course LLM engineering numpy log-log regression https://easy-explain-app.vercel.app/lesson/0wx-GpS2kns 2026-10-25 0.7 https://i.ytimg.com/vi/0wx-GpS2kns/hqdefault.jpg https://i.ytimg.com/vi/0wx-GpS2kns/hqdefault.jpg How GPT Is Trained: Next-Token Prediction Explained with Math and Code Every GPT model starts with one simple game: predict the next token. Here is how that game becomes a chatbot. https://www.youtube.com/embed/0wx-GpS2kns 838 2026-10-25T05:13:18.155Z 24 yes how GPT is trained next-token prediction language model pretraining cross-entropy loss perplexity explained autoregressive model teacher forcing causal masking LLM training data data mixture deduplication PyTorch language model character-level model generative AI course LLM engineering transformer training https://easy-explain-app.vercel.app/lesson/XJJaWGSVIgE 2026-10-25 0.7 https://i.ytimg.com/vi/XJJaWGSVIgE/hqdefault.jpg https://i.ytimg.com/vi/XJJaWGSVIgE/hqdefault.jpg The Bellman Equation: How a Gridworld Learns Its Values What if every square on a map could learn how good it is to stand there? That is the Bellman equation. https://www.youtube.com/embed/XJJaWGSVIgE 68 2026-10-25T05:13:04.070Z 267 yes bellman equation gridworld reinforcement learning value function value iteration discount factor machine learning dynamic programming markov decision process state value RL explained manim machine learning course AI basics shorts https://easy-explain-app.vercel.app/lesson/3vv61v7GAcU 2026-10-25 0.7 https://i.ytimg.com/vi/3vv61v7GAcU/hqdefault.jpg https://i.ytimg.com/vi/3vv61v7GAcU/hqdefault.jpg LoRA Explained: Fine-Tune a Giant Model by Training a Sliver What if you could fine-tune a giant model by training less than two percent of one layer? https://www.youtube.com/embed/3vv61v7GAcU 56 2026-10-25T05:12:59.139Z 466 yes LoRA low-rank adaptation fine-tuning parameter efficient fine-tuning PEFT machine learning deep learning large language models LLM low rank matrix matrix factorization transformers AI explained machine learning course Shorts https://easy-explain-app.vercel.app/lesson/7s2cOT9R9i4 2026-10-24 0.7 https://i.ytimg.com/vi/7s2cOT9R9i4/hqdefault.jpg https://i.ytimg.com/vi/7s2cOT9R9i4/hqdefault.jpg Why Attention Divides by Square Root of d Why does attention divide by the square root of d? Because without it, softmax collapses into a spike. https://www.youtube.com/embed/7s2cOT9R9i4 63 2026-10-24T06:18:44.934Z 19 yes attention mechanism scaled dot-product attention square root of d transformers explained softmax machine learning deep learning variance dot product self-attention query key vanishing gradients manim 3blue1brown style machine learning course LLM https://easy-explain-app.vercel.app/lesson/PJKP02HMNe0 2026-10-24 0.7 https://i.ytimg.com/vi/PJKP02HMNe0/hqdefault.jpg https://i.ytimg.com/vi/PJKP02HMNe0/hqdefault.jpg Why Adam Works: Momentum Plus Scaling Why does plain gradient descent zig-zag, and how does Adam glide straight to the minimum? https://www.youtube.com/embed/PJKP02HMNe0 57 2026-10-24T06:18:39.995Z 51 yes adam optimizer momentum sgd gradient descent machine learning deep learning optimizers explained neural network training manim 3blue1brown style machine learning course learning rate rmsprop optimization ai explained https://easy-explain-app.vercel.app/lesson/_RiQ4o_bjig 2026-10-24 0.7 https://i.ytimg.com/vi/_RiQ4o_bjig/hqdefault.jpg https://i.ytimg.com/vi/_RiQ4o_bjig/hqdefault.jpg The Kernel Trick: Higher Dimensions for Free Can't separate your data with a line? Lift it to a higher dimension, without ever going there. https://www.youtube.com/embed/_RiQ4o_bjig 54 2026-10-24T06:18:35.274Z 134 yes kernel trick machine learning support vector machine SVM kernel methods feature map higher dimensions machine learning for beginners dot product RBF kernel polynomial kernel linear separability Manim 3Blue1Brown style machine learning course data science https://easy-explain-app.vercel.app/lesson/zFZGoTRzlBA 2026-10-24 0.7 https://i.ytimg.com/vi/zFZGoTRzlBA/hqdefault.jpg https://i.ytimg.com/vi/zFZGoTRzlBA/hqdefault.jpg Gradient Boosting: Just Keep Fitting the Mistakes One tree is weak. Three trees that fix each other's mistakes are not. https://www.youtube.com/embed/zFZGoTRzlBA 53 2026-10-24T06:18:31.039Z 138 yes gradient boosting machine learning boosting residuals decision trees XGBoost ensemble learning negative gradient machine learning for beginners data science manim 3blue1brown style ml course supervised learning machine learning shorts https://easy-explain-app.vercel.app/lesson/HT4rbhWj3ho 2026-10-24 0.7 https://i.ytimg.com/vi/HT4rbhWj3ho/hqdefault.jpg https://i.ytimg.com/vi/HT4rbhWj3ho/hqdefault.jpg Zero Correlation Does NOT Mean Independent (Proof in 50 Seconds) Two variables can be perfectly dependent and still have a correlation of zero. https://www.youtube.com/embed/HT4rbhWj3ho 52 2026-10-24T06:18:25.382Z 18 yes correlation zero correlation independence statistics machine learning covariance parabola data science correlation vs dependence linear association pearson correlation math visualization manim machine learning course probability shorts https://easy-explain-app.vercel.app/lesson/tnag53VDTMo 2026-10-24 0.7 https://i.ytimg.com/vi/tnag53VDTMo/hqdefault.jpg https://i.ytimg.com/vi/tnag53VDTMo/hqdefault.jpg Lagrange Multipliers in 50 Seconds: Where Gradients Line Up You want the highest point, but you must stay on a path. Lagrange multipliers solve it with one picture and one equation. https://www.youtube.com/embed/tnag53VDTMo 50 2026-10-24T06:18:20.960Z 1314 yes Lagrange multipliers constrained optimization gradient contour plot tangent calculus multivariable calculus machine learning math optimization math visualization Manim 3Blue1Brown style parallel gradients machine learning course Shorts https://easy-explain-app.vercel.app/lesson/7yt2UOU4Drc 2026-10-24 0.7 https://i.ytimg.com/vi/7yt2UOU4Drc/hqdefault.jpg https://i.ytimg.com/vi/7yt2UOU4Drc/hqdefault.jpg The Chain Rule in 50 Seconds: Gears, Rates and Backprop Every neural network learns with one rule from calculus, and you can see it in 50 seconds. https://www.youtube.com/embed/7yt2UOU4Drc 52 2026-10-24T06:18:16.504Z 1374 yes chain rule chain rule explained calculus calculus for machine learning backpropagation backprop computational graph derivatives neural networks machine learning math 3blue1brown style manim math shorts learn calculus deep learning basics https://easy-explain-app.vercel.app/lesson/24e0iMnKg88 2026-10-24 0.7 https://i.ytimg.com/vi/24e0iMnKg88/hqdefault.jpg https://i.ytimg.com/vi/24e0iMnKg88/hqdefault.jpg The 99% Accuracy That Was a Lie (Data Leakage Explained) A model scored 99% accuracy and was completely useless. The cause: data leakage. https://www.youtube.com/embed/24e0iMnKg88 53 2026-10-24T06:18:12.179Z 347 yes data leakage machine learning scikit-learn pipeline train test split StandardScaler overfitting 99% accuracy machine learning mistakes data science python ml model evaluation feature leakage ML for beginners machine learning course Shorts https://easy-explain-app.vercel.app/lesson/QtvF2-JA0tQ 2026-10-24 0.7 https://i.ytimg.com/vi/QtvF2-JA0tQ/hqdefault.jpg https://i.ytimg.com/vi/QtvF2-JA0tQ/hqdefault.jpg Why Broadcasting Is NumPy's Superpower (and Its Silent Bug) One missing dimension can wreck your model, and NumPy will never warn you. https://www.youtube.com/embed/QtvF2-JA0tQ 55 2026-10-24T06:18:07.716Z 100 yes numpy broadcasting numpy broadcasting python machine learning array shapes numpy shapes reshape data science python tips linear algebra ml from scratch numpy bug vectorization manim shorts https://easy-explain-app.vercel.app/lesson/I1hRcbRTP-o 2026-10-24 0.7 https://i.ytimg.com/vi/I1hRcbRTP-o/hqdefault.jpg https://i.ytimg.com/vi/I1hRcbRTP-o/hqdefault.jpg Machine Learning Course Wrap-Up: The Full Map and Your Roadmap Beyond You finished 70 lessons of Machine Learning from Zero to Advanced. Here is the complete map of everything you learned, the six principles that tie it together, and a clear roadmap for what to do next. https://www.youtube.com/embed/I1hRcbRTP-o 818 2026-10-24T06:18:02.342Z 51 yes machine learning course machine learning roadmap ml wrap-up how to become ml engineer data scientist roadmap applied llm roadmap machine learning for beginners deep learning transformers reinforcement learning mlops kaggle practice ml portfolio paper reading what to learn after machine learning machine learning from zero to advanced https://easy-explain-app.vercel.app/lesson/FqSmDj7BpIo 2026-10-23 0.7 https://i.ytimg.com/vi/FqSmDj7BpIo/hqdefault.jpg https://i.ytimg.com/vi/FqSmDj7BpIo/hqdefault.jpg Forecasting to Production: Time Series, FastAPI, Docker and Drift Monitoring Can a forecast survive contact with the real world? In this capstone we build a time-series forecaster honestly, then ship it as a tested, containerized, monitored API. https://www.youtube.com/embed/FqSmDj7BpIo 778 2026-10-23T05:19:16.069Z 6 yes time series forecasting seasonal naive ARIMA ETS gradient boosting lag features rolling origin evaluation MASE MAPE forecast intervals data leakage FastAPI Docker MLOps drift monitoring machine learning capstone https://easy-explain-app.vercel.app/lesson/ANgBPEtDOrM 2026-10-23 0.7 https://i.ytimg.com/vi/ANgBPEtDOrM/hqdefault.jpg https://i.ytimg.com/vi/ANgBPEtDOrM/hqdefault.jpg Capstone: Train a CNN and a Tiny GPT From Scratch (Grad-CAM, Perplexity) Two real projects, one video: classify images with a CNN and transfer learning, then train a tiny GPT that writes Shakespeare-like text. https://www.youtube.com/embed/ANgBPEtDOrM 861 2026-10-23T05:19:00.830Z 8 yes machine learning capstone image classification CIFAR-10 CNN PyTorch transfer learning data augmentation confusion matrix Grad-CAM character-level language model tiny GPT nanoGPT style perplexity temperature sampling transformer from scratch deep learning project ML course PyTorch tutorial https://easy-explain-app.vercel.app/lesson/QciYJlJpU_M 2026-10-23 0.7 https://i.ytimg.com/vi/QciYJlJpU_M/hqdefault.jpg https://i.ytimg.com/vi/QciYJlJpU_M/hqdefault.jpg Customer Segmentation and a Recommender From Scratch (Capstone 7-8) Build two classic ML products from scratch: a customer segmentation with RFM features, k-means and Gaussian mixtures, and a movie recommender with item-based neighbors and matrix factorization. https://www.youtube.com/embed/QciYJlJpU_M 892 2026-10-23T05:18:47.172Z 10 yes customer segmentation RFM analysis k-means clustering gaussian mixture model silhouette score BIC model selection PCA visualization recommendation system collaborative filtering matrix factorization cosine similarity ALS SGD precision at k cold start problem machine learning capstone MovieLens python machine learning https://easy-explain-app.vercel.app/lesson/byiMkOlGXG0 2026-10-23 0.7 https://i.ytimg.com/vi/byiMkOlGXG0/hqdefault.jpg https://i.ytimg.com/vi/byiMkOlGXG0/hqdefault.jpg Spam Filter and Text Classifier From Scratch: Naive Bayes to Transformers Can a 1960s-style probability trick compete with a fine-tuned Transformer? We build a spam filter from scratch, then a multi-class text classifier, and compare them honestly. https://www.youtube.com/embed/byiMkOlGXG0 763 2026-10-23T05:18:32.618Z 8 yes spam classification naive bayes from scratch text classification tf-idf logistic regression laplace smoothing nlp tutorial fine-tuning transformer distilbert machine learning capstone error analysis data leakage label noise confusion matrix machine learning course https://easy-explain-app.vercel.app/lesson/ut6ijfBnO_Y 2026-10-23 0.7 https://i.ytimg.com/vi/ut6ijfBnO_Y/hqdefault.jpg https://i.ytimg.com/vi/ut6ijfBnO_Y/hqdefault.jpg Credit Risk, Fairness and Anomaly Detection: Capstone 3 and 4 Can a model that is accurate still be unfair? In this capstone we build a credit-risk classifier end to end, then audit it for fairness, and finally hunt rare anomalies with no labels at all. https://www.youtube.com/embed/ut6ijfBnO_Y 657 2026-10-23T05:18:19.054Z 3 yes machine learning credit risk fairness in machine learning demographic parity equalized odds model card class imbalance calibration gradient boosting logistic regression anomaly detection isolation forest mahalanobis distance autoencoder precision at k PR-AUC capstone project https://easy-explain-app.vercel.app/lesson/45G2SdpIR94 2026-10-23 0.7 https://i.ytimg.com/vi/45G2SdpIR94/hqdefault.jpg https://i.ytimg.com/vi/45G2SdpIR94/hqdefault.jpg Two Full ML Projects: House Prices and Customer Churn, Start to Finish Two end-to-end tabular machine learning projects, built with one repeatable template. https://www.youtube.com/embed/45G2SdpIR94 686 2026-10-23T05:18:00.342Z 7 yes machine learning capstone house price prediction customer churn prediction end to end machine learning project gradient boosting SHAP precision recall curve imbalanced classification model calibration residual analysis log transform target data science project scikit-learn tutorial machine learning from zero to advanced threshold selection business cost https://easy-explain-app.vercel.app/lesson/VbGnNgIIbeo 2026-10-23 0.7 https://i.ytimg.com/vi/VbGnNgIIbeo/hqdefault.jpg https://i.ytimg.com/vi/VbGnNgIIbeo/hqdefault.jpg MLOps: Reproducible Pipelines, Serving, Deployment and Monitoring Your model works in a notebook. Now what? This lesson shows how to take a machine learning model into production and keep it healthy. https://www.youtube.com/embed/VbGnNgIIbeo 816 2026-10-23T05:17:46.941Z 6 yes MLOps machine learning model deployment FastAPI Docker DVC MLflow data drift PSI KS test ONNX pickle security model monitoring canary deployment shadow deployment reproducibility machine learning course https://easy-explain-app.vercel.app/lesson/pOtWsuMS0CQ 2026-10-23 0.7 https://i.ytimg.com/vi/pOtWsuMS0CQ/hqdefault.jpg https://i.ytimg.com/vi/pOtWsuMS0CQ/hqdefault.jpg Trustworthy ML: Uncertainty, Shift, Fairness and Robustness Explained Your model scores 95 percent in testing. Why might you still not trust it? https://www.youtube.com/embed/pOtWsuMS0CQ 869 2026-10-23T05:17:28.274Z 6 yes trustworthy machine learning conformal prediction aleatoric vs epistemic uncertainty distribution shift covariate shift isolation forest permutation importance partial dependence LIME causal inference fairness in machine learning demographic parity equalized odds FGSM adversarial examples MC dropout machine learning course https://easy-explain-app.vercel.app/lesson/8UoERDhvwBc 2026-10-23 0.7 https://i.ytimg.com/vi/8UoERDhvwBc/hqdefault.jpg https://i.ytimg.com/vi/8UoERDhvwBc/hqdefault.jpg Bayesian ML: MAP, Priors, and Gaussian Processes Explained What if your model could say "I'm not sure" and mean it? Lesson 61 shows how probability turns guessing into reasoning. https://www.youtube.com/embed/8UoERDhvwBc 881 2026-10-23T05:17:13.902Z 29 yes machine learning bayesian inference MAP estimation gaussian processes bayesian linear regression beta binomial conjugate prior L2 regularization prior L1 laplace prior MCMC variational inference probabilistic machine learning graphical models naive bayes machine learning course https://easy-explain-app.vercel.app/lesson/lU4UiB3x1V0 2026-10-23 0.7 https://i.ytimg.com/vi/lU4UiB3x1V0/hqdefault.jpg https://i.ytimg.com/vi/lU4UiB3x1V0/hqdefault.jpg LLM Engineering Roadmap: What You Learned and Where to Go Next You finished the course. Now what? This final lesson turns forty lessons into one mental map, from foundations to production. https://www.youtube.com/embed/lU4UiB3x1V0 664 2026-10-23T05:16:55.801Z 46 yes LLM engineering generative AI roadmap what to learn after LLM course AI agents RAG fine-tuning prompt engineering LLM production AI portfolio projects how to read ML papers mixture of experts state space models interpretability reasoning models machine learning roadmap course wrap up https://easy-explain-app.vercel.app/lesson/su8pblpqdY4 2026-10-22 0.7 https://i.ytimg.com/vi/su8pblpqdY4/hqdefault.jpg https://i.ytimg.com/vi/su8pblpqdY4/hqdefault.jpg Reinforcement Learning II: TD, Q-Learning, Policy Gradients, DQN and PPO How does an agent learn from a single step instead of waiting for the whole episode? This lesson takes you from temporal difference learning to Q-learning, Deep Q-Networks, policy gradients and PPO, with every formula worked through by hand. https://www.youtube.com/embed/su8pblpqdY4 1030 2026-10-22T06:15:08.303Z 9 yes reinforcement learning temporal difference learning TD(0) Q-learning SARSA cliff walking epsilon greedy UCB bandit deep Q network DQN replay buffer target network policy gradient REINFORCE actor critic PPO reward hacking CartPole machine learning course Machine Learning from Zero to Advanced https://easy-explain-app.vercel.app/lesson/EmuVFna7xsw 2026-10-22 0.7 https://i.ytimg.com/vi/EmuVFna7xsw/hqdefault.jpg https://i.ytimg.com/vi/EmuVFna7xsw/hqdefault.jpg Reinforcement Learning I: MDPs, Bellman Equations, DP and Monte Carlo How does a machine learn to act when nobody tells it the right answer? Lesson 59 builds reinforcement learning from the ground up. https://www.youtube.com/embed/EmuVFna7xsw 892 2026-10-22T06:14:52.269Z 21 yes reinforcement learning markov decision process MDP bellman equation dynamic programming policy iteration value iteration monte carlo methods epsilon greedy on-policy off-policy value function discount factor gridworld machine learning course RL tutorial https://easy-explain-app.vercel.app/lesson/5-IcrmTgoro 2026-10-22 0.7 https://i.ytimg.com/vi/5-IcrmTgoro/hqdefault.jpg https://i.ytimg.com/vi/5-IcrmTgoro/hqdefault.jpg How LLMs Actually Run: Decoding, KV Cache, Latency and Limits Training gets the headlines, but inference is where language models meet reality: memory, speed, cost and mistakes. https://www.youtube.com/embed/5-IcrmTgoro 820 2026-10-22T06:14:37.519Z 19 yes LLM inference KV cache temperature softmax top-p sampling top-k sampling beam search quantization distillation batching latency vs throughput context window quadratic attention hallucination retrieval augmented generation machine learning course transformers explained https://easy-explain-app.vercel.app/lesson/YDtcMzyMsI8 2026-10-22 0.7 https://i.ytimg.com/vi/YDtcMzyMsI8/hqdefault.jpg https://i.ytimg.com/vi/YDtcMzyMsI8/hqdefault.jpg Pretraining, Fine-Tuning and LoRA Explained With Real Numbers How does a model that only predicts the next word become a helpful assistant, and how can you adapt a huge model on one GPU? https://www.youtube.com/embed/YDtcMzyMsI8 950 2026-10-22T06:14:23.106Z 15 yes pretraining fine-tuning LoRA QLoRA PEFT perplexity masked language modeling next token prediction RLHF DPO instruction tuning scaling laws Hugging Face catastrophic forgetting machine learning course transformers https://easy-explain-app.vercel.app/lesson/Zp_WLVXYeGA 2026-10-22 0.7 https://i.ytimg.com/vi/Zp_WLVXYeGA/hqdefault.jpg https://i.ytimg.com/vi/Zp_WLVXYeGA/hqdefault.jpg The Full Transformer Block: Residuals, LayerNorm, GPT vs BERT vs T5 Every large language model is one small block stacked over and over. In this lesson we assemble that block piece by piece and run it ourselves. https://www.youtube.com/embed/Zp_WLVXYeGA 948 2026-10-22T06:14:05.400Z 28 yes transformer block layer normalization residual connections feed-forward network GELU pre-LN vs post-LN causal mask cross-attention GPT architecture BERT T5 encoder decoder decoder-only PyTorch transformer parameter count machine learning course deep learning tutorial https://easy-explain-app.vercel.app/lesson/ENqhiW4D0qg 2026-10-22 0.7 https://i.ytimg.com/vi/ENqhiW4D0qg/hqdefault.jpg https://i.ytimg.com/vi/ENqhiW4D0qg/hqdefault.jpg Attention Explained With Real Numbers: Scaled Dot-Product and Multi-Head Attention is the engine of every transformer, but most explanations stop at a diagram. In this lesson we compute it by hand, number by number. https://www.youtube.com/embed/ENqhiW4D0qg 806 2026-10-22T06:13:49.032Z 30 yes attention mechanism scaled dot-product attention multi-head attention transformer self-attention queries keys values softmax causal mask attention explained machine learning course deep learning NumPy attention PyTorch attention attention heatmap why divide by sqrt d_k LLM fundamentals https://easy-explain-app.vercel.app/lesson/Iy_B-XHWHMI 2026-10-22 0.7 https://i.ytimg.com/vi/Iy_B-XHWHMI/hqdefault.jpg https://i.ytimg.com/vi/Iy_B-XHWHMI/hqdefault.jpg Tokenization and Embeddings: How Text Becomes Numbers (BPE by Hand) A language model has never read a word. It only sees integers. Here is exactly how text becomes numbers, worked out by hand. https://www.youtube.com/embed/Iy_B-XHWHMI 962 2026-10-22T06:13:34.352Z 11 yes tokenization embeddings byte pair encoding BPE word2vec skip-gram negative sampling cosine similarity positional encoding sinusoidal encoding rotary embeddings WordPiece SentencePiece Hugging Face tokenizer machine learning course NLP for beginners https://easy-explain-app.vercel.app/lesson/3ZM-ZKsSBRI 2026-10-22 0.7 https://i.ytimg.com/vi/3ZM-ZKsSBRI/hqdefault.jpg https://i.ytimg.com/vi/3ZM-ZKsSBRI/hqdefault.jpg LSTM, GRU and Backpropagation Through Time Explained With Numbers Why can a simple RNN read ten words but forget a hundred? This lesson unrolls the RNN, derives backpropagation through time, and shows the product of Jacobians behind vanishing and exploding gradients. https://www.youtube.com/embed/3ZM-ZKsSBRI 912 2026-10-22T06:13:20.104Z 18 yes LSTM GRU backpropagation through time BPTT vanishing gradient exploding gradient RNN recurrent neural network gradient clipping teacher forcing character level language model PyTorch nn.LSTM machine learning course deep learning tutorial NumPy LSTM cell state https://easy-explain-app.vercel.app/lesson/lI17nJKRu_Q 2026-10-22 0.7 https://i.ytimg.com/vi/lI17nJKRu_Q/hqdefault.jpg https://i.ytimg.com/vi/lI17nJKRu_Q/hqdefault.jpg CNN Architectures: LeNet, ResNet, Transfer Learning and Grad-CAM Explained What if you could train a great image classifier with a few hundred photos and a few minutes of compute? In this lesson we build a small CNN in PyTorch, follow the architecture story from LeNet to VGG to ResNet, and see why skip connections save gradients. https://www.youtube.com/embed/lI17nJKRu_Q 881 2026-10-22T06:13:05.873Z 4 yes cnn architectures resnet explained transfer learning pytorch cnn tutorial grad-cam skip connections vgg lenet data augmentation fine-tuning feature extraction confusion matrix machine learning course deep learning for beginners visualizing cnn features vanishing gradients https://easy-explain-app.vercel.app/lesson/JcH2-HBVRyk 2026-10-22 0.7 https://i.ytimg.com/vi/JcH2-HBVRyk/hqdefault.jpg https://i.ytimg.com/vi/JcH2-HBVRyk/hqdefault.jpg Convolution Explained: Kernels, Stride, Padding and Pooling by Hand A 150 million weight layer or 1,792 weights? Learn why convolution is the right tool for images. https://www.youtube.com/embed/JcH2-HBVRyk 839 2026-10-22T06:12:51.203Z 17 yes convolution convolutional neural network CNN explained kernel stride padding pooling max pooling receptive field 1x1 convolution conv2d from scratch NumPy PyTorch deep learning machine learning course cross-correlation https://easy-explain-app.vercel.app/lesson/VY9uxf-iL24 2026-10-21 0.7 https://i.ytimg.com/vi/VY9uxf-iL24/hqdefault.jpg https://i.ytimg.com/vi/VY9uxf-iL24/hqdefault.jpg PyTorch Training Loops, DataLoaders, GPUs and Debugging Explained Your model trains, but is it trained correctly? Learn the full PyTorch workflow, from Dataset and DataLoader to a clean training loop, GPU use and a practical debugging checklist. https://www.youtube.com/embed/VY9uxf-iL24 774 2026-10-21T05:15:01.697Z 12 yes pytorch pytorch tutorial training loop dataloader dataset nn.Module torch.optim gpu training mixed precision autocast early stopping checkpointing debugging neural networks overfit one batch reproducibility machine learning course deep learning https://easy-explain-app.vercel.app/lesson/bFJjj6ImHj4 2026-10-21 0.7 https://i.ytimg.com/vi/bFJjj6ImHj4/hqdefault.jpg https://i.ytimg.com/vi/bFJjj6ImHj4/hqdefault.jpg PyTorch Tensors and Autograd: Gradients You Can Verify by Hand What if you never had to differentiate by hand again, but could still prove the answer is right? https://www.youtube.com/embed/bFJjj6ImHj4 873 2026-10-21T05:14:44.004Z 7 yes pytorch pytorch tutorial autograd pytorch tensors backpropagation computational graph requires_grad zero_grad torch.no_grad detach machine learning course deep learning linear regression pytorch finite difference gradient check custom autograd function numpy vs pytorch broadcasting https://easy-explain-app.vercel.app/lesson/gognHqfQcWs 2026-10-21 0.7 https://i.ytimg.com/vi/gognHqfQcWs/hqdefault.jpg https://i.ytimg.com/vi/gognHqfQcWs/hqdefault.jpg Dropout and Batch Normalization Explained With Numbers and NumPy Code Why does randomly deleting neurons make a network better? And how does normalizing activations make training faster? We derive dropout and batch normalization step by step with concrete numbers, including the full backward pass. https://www.youtube.com/embed/gognHqfQcWs 971 2026-10-21T05:14:28.363Z 61 yes dropout batch normalization batchnorm regularization deep learning neural networks inverted dropout layer normalization AdamW weight decay early stopping data augmentation backpropagation machine learning course NumPy from scratch gradient check https://easy-explain-app.vercel.app/lesson/TSo1n2doPzE 2026-10-21 0.7 https://i.ytimg.com/vi/TSo1n2doPzE/hqdefault.jpg https://i.ytimg.com/vi/TSo1n2doPzE/hqdefault.jpg Weight Initialization, Vanishing Gradients and Optimizers: Why Deep Nets Train Why does a deep network sometimes refuse to learn at all? Often the answer is decided before the first training step. https://www.youtube.com/embed/TSo1n2doPzE 895 2026-10-21T05:14:13.081Z 6 yes weight initialization xavier initialization glorot initialization he initialization vanishing gradients exploding gradients gradient clipping adam optimizer momentum nesterov rmsprop learning rate schedule learning rate warmup deep learning tutorial machine learning course optimizers explained https://easy-explain-app.vercel.app/lesson/SWAlHphGh8E 2026-10-21 0.7 https://i.ytimg.com/vi/SWAlHphGh8E/hqdefault.jpg https://i.ytimg.com/vi/SWAlHphGh8E/hqdefault.jpg ARIMA and Exponential Smoothing Explained With Real Numbers Two classic forecasting families, ARIMA and exponential smoothing, built from the ground up with hand-worked numbers. https://www.youtube.com/embed/SWAlHphGh8E 948 2026-10-21T05:13:58.861Z 19 yes ARIMA exponential smoothing time series forecasting Holt-Winters SARIMA ACF PACF ADF test Yule-Walker Ljung-Box statsmodels machine learning course stationarity differencing lag operator rolling origin evaluation forecast intervals https://easy-explain-app.vercel.app/lesson/ZOVWaRW4rOU 2026-10-21 0.7 https://i.ytimg.com/vi/ZOVWaRW4rOU/hqdefault.jpg https://i.ytimg.com/vi/ZOVWaRW4rOU/hqdefault.jpg Hyperparameter Search and Bayesian Optimization Explained With Numbers Stop guessing learning rates. Learn how to tune models with a plan: random search, nested cross-validation, Hyperband, and Bayesian optimization. https://www.youtube.com/embed/ZOVWaRW4rOU 866 2026-10-21T05:13:45.494Z 9 yes hyperparameter tuning hyperparameter search bayesian optimization random search vs grid search expected improvement optuna tutorial nested cross validation successive halving hyperband gaussian process surrogate machine learning course machine learning from scratch acquisition function validation overfitting learning rate log scale python machine learning https://easy-explain-app.vercel.app/lesson/brnKyvuAj1s 2026-10-21 0.7 https://i.ytimg.com/vi/brnKyvuAj1s/hqdefault.jpg https://i.ytimg.com/vi/brnKyvuAj1s/hqdefault.jpg Expectation-Maximization for Gaussian Mixtures: The Full Math and Code What do you do when the cluster label of every data point is hidden? Expectation-Maximization is the elegant answer, and in this lesson we build it from the ground up. https://www.youtube.com/embed/brnKyvuAj1s 970 2026-10-21T05:13:30.982Z 11 yes expectation maximization EM algorithm gaussian mixture model GMM machine learning responsibilities ELBO Jensen inequality k-means BIC AIC latent variable NumPy from scratch scikit-learn GaussianMixture unsupervised learning machine learning course soft clustering https://easy-explain-app.vercel.app/lesson/b45BFw0izZI 2026-10-21 0.7 https://i.ytimg.com/vi/b45BFw0izZI/hqdefault.jpg https://i.ytimg.com/vi/b45BFw0izZI/hqdefault.jpg The SVM Dual and the Kernel Trick: The Math Explained Step by Step How can a machine learning model work in infinite dimensions without ever computing a single coordinate? The answer is the SVM dual and the kernel trick. https://www.youtube.com/embed/b45BFw0izZI 942 2026-10-21T05:13:14.914Z 34 yes svm dual kernel trick support vector machine lagrangian KKT conditions complementary slackness soft margin svm hinge loss polynomial kernel rbf kernel mercer condition support vector regression epsilon tube machine learning math machine learning course sklearn SVC dual problem https://easy-explain-app.vercel.app/lesson/69WmnK-_Zlk 2026-10-21 0.7 https://i.ytimg.com/vi/69WmnK-_Zlk/hqdefault.jpg https://i.ytimg.com/vi/69WmnK-_Zlk/hqdefault.jpg Gradient Boosting and XGBoost: The Math, Derived Step by Step Why do boosted trees win so many tabular data problems? Because they are gradient descent, performed on functions instead of numbers. https://www.youtube.com/embed/69WmnK-_Zlk 892 2026-10-21T05:13:01.366Z 13 yes gradient boosting xgboost xgboost math machine learning course pseudo-residuals gradient descent in function space lightgbm catboost decision trees boosting explained taylor expansion hessian early stopping hyperparameter tuning tabular data ml from scratch https://easy-explain-app.vercel.app/lesson/3aumzTBOzbQ 2026-10-21 0.7 https://i.ytimg.com/vi/3aumzTBOzbQ/hqdefault.jpg https://i.ytimg.com/vi/3aumzTBOzbQ/hqdefault.jpg Logistic Regression Gradient Derived From Scratch (Full Math + NumPy) One line of algebra hides inside every classifier you have ever trained: gradient equals X transpose times prediction minus label. In this lesson we derive it from scratch. https://www.youtube.com/embed/3aumzTBOzbQ 896 2026-10-21T05:12:44.061Z 17 yes logistic regression logistic regression gradient binary cross entropy sigmoid derivative softmax gradient newton method IRLS hessian perfect separation log-sum-exp machine learning from scratch numpy tutorial maximum likelihood chain rule scikit-learn machine learning math https://easy-explain-app.vercel.app/lesson/J9tqB8iGkP8 2026-10-20 0.7 https://i.ytimg.com/vi/J9tqB8iGkP8/hqdefault.jpg https://i.ytimg.com/vi/J9tqB8iGkP8/hqdefault.jpg Linear Regression From Scratch in NumPy: Assumptions and Residual Analysis Your linear regression runs and gives numbers. But can you trust them? https://www.youtube.com/embed/J9tqB8iGkP8 781 2026-10-20T06:20:07.196Z 10 yes linear regression linear regression from scratch numpy residual analysis regression assumptions homoscedasticity Q-Q plot Cook's distance leverage R squared adjusted R squared standard errors multicollinearity normal equation machine learning course QR decomposition https://easy-explain-app.vercel.app/lesson/gz8iIMTJDqY 2026-10-20 0.7 https://i.ytimg.com/vi/gz8iIMTJDqY/hqdefault.jpg https://i.ytimg.com/vi/gz8iIMTJDqY/hqdefault.jpg Confidence Intervals and Hypothesis Tests for ML: Is Your Model Really Better? Your new model scores 90 percent and the old one scores 88. Is that a real improvement, or just luck? https://www.youtube.com/embed/gz8iIMTJDqY 797 2026-10-20T06:19:51.183Z 15 yes machine learning statistics for machine learning confidence interval hypothesis testing p-value standard error bootstrap t-test chi-square test A/B testing type I error statistical power p-hacking multiple comparisons sampling distribution central limit theorem model evaluation https://easy-explain-app.vercel.app/lesson/dHQit23qfN0 2026-10-20 0.7 https://i.ytimg.com/vi/dHQit23qfN0/hqdefault.jpg https://i.ytimg.com/vi/dHQit23qfN0/hqdefault.jpg Covariance, Correlation and the Multivariate Gaussian Explained Two numbers can move together, yet a correlation of zero can still hide a perfect relationship. Here is how to read covariance, correlation and the Gaussian properly. https://www.youtube.com/embed/dHQit23qfN0 937 2026-10-20T06:19:33.774Z 33 yes covariance correlation multivariate gaussian multivariate normal mahalanobis distance covariance matrix pearson correlation spearman correlation simpson's paradox correlation vs causation machine learning course machine learning math positive semi-definite covariance ellipse statistics for machine learning PCA prerequisites https://easy-explain-app.vercel.app/lesson/LdScZx8zULU 2026-10-20 0.7 https://i.ytimg.com/vi/LdScZx8zULU/hqdefault.jpg https://i.ytimg.com/vi/LdScZx8zULU/hqdefault.jpg Convexity and Lagrange Multipliers: Optimization With Rules (ML Lesson 37) What if the best answer is not allowed? Learn why convex functions have no bad valleys, and how Lagrange multipliers and KKT conditions solve optimization problems with constraints. https://www.youtube.com/embed/LdScZx8zULU 842 2026-10-20T06:19:16.784Z 28 yes convexity convex optimization lagrange multipliers KKT conditions constrained optimization machine learning math hessian positive definite complementary slackness lagrangian duality SVM machine learning course gradient descent local minimum global minimum MSE convex logistic loss convex https://easy-explain-app.vercel.app/lesson/78jRUEZ09uQ 2026-10-20 0.7 https://i.ytimg.com/vi/78jRUEZ09uQ/hqdefault.jpg https://i.ytimg.com/vi/78jRUEZ09uQ/hqdefault.jpg The Chain Rule and Computational Graphs: How Gradients Really Flow Every neural network trains because of one idea from calculus: the chain rule. In this lesson we derive it from first principles, check it with real numbers, and then turn it into a machine. https://www.youtube.com/embed/78jRUEZ09uQ 1021 2026-10-20T06:18:59.396Z 35 yes chain rule computational graph backpropagation autograd reverse mode differentiation machine learning math calculus for machine learning gradient Jacobian vector-Jacobian product finite difference gradient check micrograd deep learning basics machine learning course derivative https://easy-explain-app.vercel.app/lesson/5LFFz4UvfYc 2026-10-20 0.7 https://i.ytimg.com/vi/5LFFz4UvfYc/hqdefault.jpg https://i.ytimg.com/vi/5LFFz4UvfYc/hqdefault.jpg Why Your 99% Model Is a Lie: Leakage-Safe Splits and Pipelines Your model scores 99 percent in testing and fails in production. The cause is almost always data leakage. https://www.youtube.com/embed/5LFFz4UvfYc 751 2026-10-20T06:18:41.662Z 48 yes data leakage machine learning train test split scikit-learn pipeline ColumnTransformer stratified split validation set test set target leakage preprocessing leakage group leakage temporal leakage cross validation python machine learning ML course leakage audit https://easy-explain-app.vercel.app/lesson/FXraVViHNMg 2026-10-20 0.7 https://i.ytimg.com/vi/FXraVViHNMg/hqdefault.jpg https://i.ytimg.com/vi/FXraVViHNMg/hqdefault.jpg Data Cleaning for ML: Missing Values, Outliers and Feature Types Your model is only as good as the data you feed it. Here is how to clean that data without fooling yourself. https://www.youtube.com/embed/FXraVViHNMg 821 2026-10-20T06:18:25.249Z 67 yes data cleaning missing values outliers imputation machine learning course scikit-learn SimpleImputer KNNImputer MCAR MAR MNAR IQR rule z-score winsorizing feature types data preprocessing machine learning from zero to advanced python data science https://easy-explain-app.vercel.app/lesson/0QpZwWO2OVc 2026-10-20 0.7 https://i.ytimg.com/vi/0QpZwWO2OVc/hqdefault.jpg https://i.ytimg.com/vi/0QpZwWO2OVc/hqdefault.jpg Plotting and EDA: Look at Your Data Before You Model It Four datasets can share the same summary numbers and look completely different. Learn why you must plot your data before training any model. https://www.youtube.com/embed/0QpZwWO2OVc 792 2026-10-20T06:18:08.192Z 41 yes exploratory data analysis EDA matplotlib tutorial seaborn tutorial data visualization python machine learning course pair plot correlation heatmap histogram vs KDE box plot violin plot Simpson's paradox penguins dataset overplotting misleading charts python for machine learning data science for beginners https://easy-explain-app.vercel.app/lesson/BeFK-uDJq04 2026-10-20 0.7 https://i.ytimg.com/vi/BeFK-uDJq04/hqdefault.jpg https://i.ytimg.com/vi/BeFK-uDJq04/hqdefault.jpg Pandas Essentials for Machine Learning Data (Lesson 32) Most machine learning time is spent on tables, not models. This lesson teaches the pandas skills you need, checked by hand on a tiny toy table. https://www.youtube.com/embed/BeFK-uDJq04 758 2026-10-20T06:17:53.168Z 20 yes pandas tutorial pandas for machine learning DataFrame groupby merge join pandas loc vs iloc SettingWithCopyWarning pivot table pandas design matrix numpy to pandas python data science machine learning course read_csv vectorization pandas memory optimization ML from scratch https://easy-explain-app.vercel.app/lesson/c06DC_PeDyM 2026-10-20 0.7 https://i.ytimg.com/vi/c06DC_PeDyM/hqdefault.jpg https://i.ytimg.com/vi/c06DC_PeDyM/hqdefault.jpg NumPy Arrays, Shapes and Broadcasting Explained Visually (ML Lesson 31) One million numbers, one slow Python loop, one fast NumPy line. Why is the gap so big, and how do you think in arrays? https://www.youtube.com/embed/c06DC_PeDyM 770 2026-10-20T06:17:36.978Z 18 yes numpy tutorial numpy broadcasting numpy arrays numpy shapes numpy axis numpy reshape numpy views vs copies numpy einsum vectorization python python for machine learning machine learning course machine learning from scratch boolean masking numpy numpy for beginners ndarray dtype https://easy-explain-app.vercel.app/lesson/I1EFRANihEc 2026-10-19 0.7 https://i.ytimg.com/vi/I1EFRANihEc/hqdefault.jpg https://i.ytimg.com/vi/I1EFRANihEc/hqdefault.jpg How AI Decides Which Words Matter: Attention Explained How does a model know which word in a sentence to focus on? Attention turns word similarity into weights. https://www.youtube.com/embed/I1EFRANihEc 43 2026-10-19T05:13:25.139Z 22 yes attention mechanism self attention transformers softmax query key value machine learning deep learning neural networks how attention works AI explained NLP tokens embeddings attention heatmap machine learning course shorts https://easy-explain-app.vercel.app/lesson/OwqzX5FV5HE 2026-10-19 0.7 https://i.ytimg.com/vi/OwqzX5FV5HE/hqdefault.jpg https://i.ytimg.com/vi/OwqzX5FV5HE/hqdefault.jpg Why 99% Accuracy Can Be Totally Useless A model with 99 percent accuracy that catches zero fraud? Here is how accuracy fools you on imbalanced data, and how a confusion matrix exposes it. https://www.youtube.com/embed/OwqzX5FV5HE 56 2026-10-19T05:13:20.373Z 23 yes machine learning accuracy imbalanced data fraud detection confusion matrix recall precision F1 score class imbalance data science ML for beginners model evaluation machine learning course 3blue1brown style Shorts https://easy-explain-app.vercel.app/lesson/-vV2IRSPwRo 2026-10-19 0.7 https://i.ytimg.com/vi/-vV2IRSPwRo/hqdefault.jpg https://i.ytimg.com/vi/-vV2IRSPwRo/hqdefault.jpg Why Lasso Sets Weights Exactly to Zero (Diamond vs Circle) Why does Lasso zero out coefficients while Ridge only shrinks them? One picture explains it: a diamond has corners, a circle does not. https://www.youtube.com/embed/-vV2IRSPwRo 48 2026-10-19T05:13:15.438Z 39 yes lasso regression l1 regularization ridge regression l2 regularization sparsity feature selection machine learning regularization geometry manim 3blue1brown style machine learning for beginners data science shorts l1 vs l2 constraint region https://easy-explain-app.vercel.app/lesson/4QkCeP4uOtw 2026-10-19 0.7 https://i.ytimg.com/vi/4QkCeP4uOtw/hqdefault.jpg https://i.ytimg.com/vi/4QkCeP4uOtw/hqdefault.jpg Learning Rate: Too Small, Just Right, Too Big One number decides whether your model learns or explodes. Watch three balls roll down the same loss landscape with three different learning rates. https://www.youtube.com/embed/4QkCeP4uOtw 52 2026-10-19T05:13:09.771Z 36 yes learning rate gradient descent machine learning machine learning for beginners optimization loss function contour plot divergence convergence hyperparameters deep learning neural networks 3blue1brown style manim shorts ML explained https://easy-explain-app.vercel.app/lesson/aV-k2NoEZ1s 2026-10-19 0.7 https://i.ytimg.com/vi/aV-k2NoEZ1s/hqdefault.jpg https://i.ytimg.com/vi/aV-k2NoEZ1s/hqdefault.jpg 99% Accurate Test, Positive Result: Only 50% Sick? A 99% accurate test says you're positive. So you're 99% likely sick, right? Wrong. Here's the base-rate trap in Bayes' theorem, shown by counting 10,000 people. https://www.youtube.com/embed/aV-k2NoEZ1s 49 2026-10-19T05:13:03.237Z 17 yes bayes theorem base rate fallacy base rate trap conditional probability probability medical test accuracy false positive machine learning math statistics 3blue1brown style manim math shorts bayesian reasoning precision and recall data science https://easy-explain-app.vercel.app/lesson/KQrOHLLq3T8 2026-10-19 0.7 https://i.ytimg.com/vi/KQrOHLLq3T8/hqdefault.jpg https://i.ytimg.com/vi/KQrOHLLq3T8/hqdefault.jpg Eigenvectors: The One Direction a Shear Can't Turn Almost every arrow turns when you transform space. Eigenvectors are the exception. https://www.youtube.com/embed/KQrOHLLq3T8 49 2026-10-19T05:12:58.613Z 1477 yes eigenvectors eigenvalues linear algebra shear transform PCA principal component analysis machine learning math 3blue1brown style manim matrix transformation covariance matrix math for machine learning linear algebra visualized shorts learn machine learning https://easy-explain-app.vercel.app/lesson/r5CCKbBHODg 2026-10-19 0.7 https://i.ytimg.com/vi/r5CCKbBHODg/hqdefault.jpg https://i.ytimg.com/vi/r5CCKbBHODg/hqdefault.jpg A Dot Product Is Just Alignment, Explained Visually The dot product isn't just multiply and add. It measures how much two arrows point the same way. https://www.youtube.com/embed/r5CCKbBHODg 40 2026-10-19T05:12:52.715Z 1625 yes dot product dot product explained linear algebra linear algebra for machine learning cosine similarity embeddings vector projection machine learning math 3blue1brown style manim vector alignment semantic search math shorts ML for beginners orthogonal vectors https://easy-explain-app.vercel.app/lesson/sOnTTY1zmAw 2026-10-19 0.7 https://i.ytimg.com/vi/sOnTTY1zmAw/hqdefault.jpg https://i.ytimg.com/vi/sOnTTY1zmAw/hqdefault.jpg Why Machine Learning Squares Its Mistakes (MSE by Hand) Why does machine learning square its errors? Work out mean squared error by hand with just three data points. https://www.youtube.com/embed/sOnTTY1zmAw 53 2026-10-19T05:12:47.240Z 748 yes machine learning mean squared error MSE loss function squared error gradient descent ML for beginners data science regression math for machine learning 3blue1brown style manim AI explained linear regression shorts https://easy-explain-app.vercel.app/lesson/GQSqL3a3tOU 2026-10-19 0.7 https://i.ytimg.com/vi/GQSqL3a3tOU/hqdefault.jpg https://i.ytimg.com/vi/GQSqL3a3tOU/hqdefault.jpg Reinforcement Learning, Causal Inference and MLOps: The ML Capstone What happens when a model has to learn by acting, and then survive in the real world? This final episode ties the whole course together. https://www.youtube.com/embed/GQSqL3a3tOU 984 2026-10-19T05:12:42.604Z 29 yes machine learning reinforcement learning q-learning bellman equation markov decision process policy gradient DQN gaussian process conformal prediction causal inference distribution shift anomaly detection ML fairness MLOps model drift ML capstone projects machine learning roadmap https://easy-explain-app.vercel.app/lesson/5RTu_dLI92Q 2026-10-19 0.7 https://i.ytimg.com/vi/5RTu_dLI92Q/hqdefault.jpg https://i.ytimg.com/vi/5RTu_dLI92Q/hqdefault.jpg Transformers Explained: Attention, Math and How LLMs Really Work Every modern language model is built from one idea: attention. In this episode we build a transformer from the ground up, with real numbers and visuals. https://www.youtube.com/embed/5RTu_dLI92Q 1019 2026-10-19T05:12:27.571Z 100 yes transformers attention is all you need self attention large language models LLM explained scaled dot product attention multi-head attention positional encoding tokenization byte pair encoding embeddings LoRA fine-tuning KV cache machine learning course deep learning GPT BERT layer normalization causal mask https://easy-explain-app.vercel.app/lesson/YZg2ugwLMaQ 2026-10-13 0.7 https://i.ytimg.com/vi/YZg2ugwLMaQ/hqdefault.jpg https://i.ytimg.com/vi/YZg2ugwLMaQ/hqdefault.jpg RNNs, LSTMs and Attention Explained With Real Numbers How does a neural network read a sentence one word at a time and still remember the beginning? In this episode we build recurrent networks from scratch, with real numbers at every step. https://www.youtube.com/embed/YZg2ugwLMaQ 1017 2026-10-13T18:29:45.181Z 23 yes rnn lstm gru attention mechanism sequence models recurrent neural network backpropagation through time vanishing gradient machine learning course deep learning language modeling teacher forcing padding and masking seq2seq machine learning from zero neural networks explained https://easy-explain-app.vercel.app/lesson/RXIt1OF8dYs 2026-10-13 0.7 https://i.ytimg.com/vi/RXIt1OF8dYs/hqdefault.jpg https://i.ytimg.com/vi/RXIt1OF8dYs/hqdefault.jpg PyTorch and Convolutional Neural Networks: Tensors to Transfer Learning Ever wondered how a computer actually sees an image, and how a few lines of PyTorch can teach it to recognize one? In this episode we go from tensors and autograd to convolution computed by hand, then build, train and improve a real CNN. https://www.youtube.com/embed/RXIt1OF8dYs 1182 2026-10-13T18:29:28.233Z 100 yes PyTorch PyTorch tutorial convolutional neural network CNN explained deep learning machine learning course autograd DataLoader training loop convolution by hand stride and padding pooling receptive field transfer learning data augmentation computer vision visualizing filters learn machine learning neural networks https://easy-explain-app.vercel.app/lesson/6nbA8-8SlrQ 2026-10-13 0.7 https://i.ytimg.com/vi/6nbA8-8SlrQ/hqdefault.jpg https://i.ytimg.com/vi/6nbA8-8SlrQ/hqdefault.jpg Neural Networks From Scratch: Backpropagation Derived Step by Step Build a neural network from scratch in NumPy and derive backpropagation layer by layer, with real numbers at every step. https://www.youtube.com/embed/6nbA8-8SlrQ 1074 2026-10-13T18:29:12.108Z 28 yes neural networks from scratch backpropagation explained backpropagation derivation machine learning course deep learning for beginners numpy neural network perceptron activation functions relu sigmoid tanh vanishing gradient exploding gradients he initialization xavier initialization adam optimizer dropout batch normalization gradient check chain rule neural network mini-batch gradient descent forward propagation https://easy-explain-app.vercel.app/lesson/IOQ8selTgqc 2026-10-13 0.7 https://i.ytimg.com/vi/IOQ8selTgqc/hqdefault.jpg https://i.ytimg.com/vi/IOQ8selTgqc/hqdefault.jpg Time Series and Forecasting: ARIMA, Smoothing and Leakage Explained Why does shuffling your data ruin a forecast? Learn how time series really work, step by step, with numbers you can check by hand. https://www.youtube.com/embed/IOQ8selTgqc 1179 2026-10-13T18:28:53.163Z 24 yes time series forecasting ARIMA autocorrelation stationarity differencing exponential smoothing moving average seasonality trend decomposition lag features time series cross validation MAPE MAE data leakage machine learning course Holt-Winters https://easy-explain-app.vercel.app/lesson/QtrXGxCCndg 2026-10-13 0.7 https://i.ytimg.com/vi/QtrXGxCCndg/hqdefault.jpg https://i.ytimg.com/vi/QtrXGxCCndg/hqdefault.jpg Model Evaluation and Selection Done Right: CV, Tuning, Leakage Your model scored 95 percent. Can you trust it? Probably not yet. https://www.youtube.com/embed/QtrXGxCCndg 826 2026-10-13T18:28:37.796Z 10 yes model evaluation model selection cross-validation k-fold stratified k-fold group k-fold time series split hyperparameter tuning grid search random search bayesian optimization learning curve validation curve bootstrap confidence interval data leakage nested cross-validation scikit-learn machine learning course https://easy-explain-app.vercel.app/lesson/jFOKSPLdYCc 2026-10-13 0.7 https://i.ytimg.com/vi/jFOKSPLdYCc/hqdefault.jpg https://i.ytimg.com/vi/jFOKSPLdYCc/hqdefault.jpg PCA, SVD, t-SNE and UMAP: Dimensionality Reduction Explained Visually How do you squeeze a thousand features into two without losing what matters? This video builds PCA from first principles, derives it with a Lagrange multiplier, connects it to SVD, then explains t-SNE and UMAP and how not to misread their pictures. https://www.youtube.com/embed/jFOKSPLdYCc 989 2026-10-13T18:28:22.075Z 45 yes PCA principal component analysis SVD singular value decomposition t-SNE UMAP dimensionality reduction machine learning explained variance whitening covariance matrix eigenvectors Lagrange multiplier curse of dimensionality data visualization machine learning course NumPy PCA from scratch https://easy-explain-app.vercel.app/lesson/2tFni5w6xbY 2026-10-13 0.7 https://i.ytimg.com/vi/2tFni5w6xbY/hqdefault.jpg https://i.ytimg.com/vi/2tFni5w6xbY/hqdefault.jpg Clustering Explained: k-Means, Hierarchical, DBSCAN and GMM How do you find groups in data when nobody tells you the answer? In this episode we build clustering from the ground up, with math, worked numbers and animation. https://www.youtube.com/embed/2tFni5w6xbY 1112 2026-10-13T18:28:07.772Z 19 yes clustering k-means k-means++ hierarchical clustering dendrogram DBSCAN gaussian mixture model expectation maximization silhouette score elbow method unsupervised learning machine learning tutorial machine learning course Lloyd's algorithm cluster evaluation https://easy-explain-app.vercel.app/lesson/N7cNfErHgS0 2026-10-13 0.7 https://i.ytimg.com/vi/N7cNfErHgS0/hqdefault.jpg https://i.ytimg.com/vi/N7cNfErHgS0/hqdefault.jpg Support Vector Machines and Kernels Explained: Margins, Duals, RBF What if the best classifier is the one that stays as far from the data as possible? In this episode we build Support Vector Machines from the ground up. https://www.youtube.com/embed/N7cNfErHgS0 1091 2026-10-13T18:27:49.963Z 30 yes support vector machine SVM kernel trick RBF kernel polynomial kernel hinge loss soft margin hard margin Lagrange multipliers dual form SVR support vector regression scikit-learn SVC C and gamma tuning machine learning course maximum margin classifier KKT conditions machine learning from scratch https://easy-explain-app.vercel.app/lesson/5EvGJ8-1NHY 2026-10-13 0.7 https://i.ytimg.com/vi/5EvGJ8-1NHY/hqdefault.jpg https://i.ytimg.com/vi/5EvGJ8-1NHY/hqdefault.jpg Random Forests, Boosting and SHAP Explained With Real Numbers Why does a crowd of weak decision trees beat one clever model? This episode builds bagging, random forests, AdaBoost, gradient boosting, XGBoost, stacking and SHAP from the ground up, with hand-worked numbers. https://www.youtube.com/embed/5EvGJ8-1NHY 980 2026-10-13T18:27:32.601Z 24 yes random forest gradient boosting adaboost xgboost lightgbm catboost shap shapley values bagging bootstrap ensemble learning stacking out-of-bag feature importance machine learning tutorial machine learning course https://easy-explain-app.vercel.app/lesson/k-ZNVu4GNMs 2026-10-13 0.7 https://i.ytimg.com/vi/k-ZNVu4GNMs/hqdefault.jpg https://i.ytimg.com/vi/k-ZNVu4GNMs/hqdefault.jpg Decision Trees Explained: Entropy, Gini, Pruning and Code From Scratch How does a machine learn a flowchart from data? In this episode we build a decision tree by hand, with real numbers, from the very first split to the final leaf. https://www.youtube.com/embed/k-ZNVu4GNMs 927 2026-10-13T18:27:18.334Z 36 yes decision trees decision tree machine learning entropy information gain gini impurity regression trees variance reduction pruning overfitting feature importance scikit-learn decision tree decision tree from scratch machine learning tutorial machine learning course CART algorithm recursive partitioning https://easy-explain-app.vercel.app/lesson/H7-1wi9Nh8A 2026-10-12 0.7 https://i.ytimg.com/vi/H7-1wi9Nh8A/hqdefault.jpg https://i.ytimg.com/vi/H7-1wi9Nh8A/hqdefault.jpg Classification Metrics, Thresholds and Calibration Explained Visually A model with 98 percent accuracy can be completely useless. In this episode we build classification metrics from scratch and learn when to trust each one. https://www.youtube.com/embed/H7-1wi9Nh8A 1036 2026-10-12T11:20:56.000Z 75 yes machine learning classification metrics confusion matrix precision and recall F1 score ROC curve AUC precision recall curve class imbalance threshold selection calibration calibration curve Brier score scikit-learn logistic regression machine learning course data science https://easy-explain-app.vercel.app/lesson/_e4mvvr9lkc 2026-10-12 0.7 https://i.ytimg.com/vi/_e4mvvr9lkc/hqdefault.jpg https://i.ytimg.com/vi/_e4mvvr9lkc/hqdefault.jpg Logistic Regression and Softmax: Derived from Scratch, Step by Step How does a model turn a straight line into a probability? In this episode we derive logistic regression and softmax from first principles, with every step worked out in numbers. https://www.youtube.com/embed/_e4mvvr9lkc 1163 2026-10-12T11:20:44.000Z 41 yes logistic regression softmax sigmoid function binary cross-entropy log odds maximum likelihood gradient descent machine learning course multiclass classification one vs rest decision boundary scikit-learn numpy from scratch machine learning for beginners classification cross entropy loss https://easy-explain-app.vercel.app/lesson/Jy5lkk6zkiA 2026-10-12 0.7 https://i.ytimg.com/vi/Jy5lkk6zkiA/hqdefault.jpg https://i.ytimg.com/vi/Jy5lkk6zkiA/hqdefault.jpg Distributions, Expectation and the Central Limit Theorem Explained Visually Why does the bell curve show up everywhere, even when the data itself is not bell shaped? In this episode we build the answer from the ground up. https://www.youtube.com/embed/Jy5lkk6zkiA 1165 2026-10-12T11:20:33.000Z 73 yes machine learning probability distributions central limit theorem expectation and variance normal distribution binomial distribution poisson distribution law of large numbers confidence intervals hypothesis testing p-value random variables PDF CDF PMF statistics for machine learning math for machine learning joint distribution manim https://easy-explain-app.vercel.app/lesson/g3J4jvnm8Bk 2026-10-12 0.7 https://i.ytimg.com/vi/g3J4jvnm8Bk/hqdefault.jpg https://i.ytimg.com/vi/g3J4jvnm8Bk/hqdefault.jpg k-Nearest Neighbors and Naive Bayes Explained: Distance, Bayes, Spam Filters What if the best classifier were just a lookup, and the other one a clever use of counting? https://www.youtube.com/embed/g3J4jvnm8Bk 1005 2026-10-12T11:20:16.000Z 42 yes k-nearest neighbors knn naive bayes machine learning machine learning course bayes theorem spam filter laplace smoothing curse of dimensionality euclidean distance manhattan distance minkowski distance gaussian naive bayes multinomial naive bayes scikit-learn python machine learning classification algorithms https://easy-explain-app.vercel.app/lesson/CkyQHRN9-Q4 2026-10-12 0.7 https://i.ytimg.com/vi/CkyQHRN9-Q4/hqdefault.jpg https://i.ytimg.com/vi/CkyQHRN9-Q4/hqdefault.jpg Linear Regression Explained: Normal Equation to Gradient Descent Two ways to fit the same line: solve it in one shot with the normal equation, or walk downhill with gradient descent. Which one should you trust? https://www.youtube.com/embed/CkyQHRN9-Q4 1002 2026-10-12T10:50:56.000Z 19 yes linear regression normal equation gradient descent machine learning tutorial least squares MSE MAE RMSE matrix calculus numpy from scratch scikit-learn feature scaling polynomial regression residual analysis machine learning course multiple linear regression ordinary least squares https://easy-explain-app.vercel.app/lesson/8ayFobAlFks 2026-10-12 0.7 https://i.ytimg.com/vi/8ayFobAlFks/hqdefault.jpg https://i.ytimg.com/vi/8ayFobAlFks/hqdefault.jpg Overfitting and the Bias-Variance Trade-off, Derived Step by Step Why does a model that scores perfectly on training data fail on new data? In this episode we build the answer from scratch: polynomial fits, training versus validation curves, and a full step-by-step derivation of the bias-variance decomposition. https://www.youtube.com/embed/8ayFobAlFks 988 2026-10-12T10:50:38.000Z 15 yes machine learning overfitting underfitting bias variance tradeoff generalization model complexity polynomial regression validation error training error bias variance decomposition data leakage baseline model machine learning course machine learning tutorial python simulation expected squared error train test split cross validation ML for beginners https://easy-explain-app.vercel.app/lesson/lxQokWFy5Qc 2026-10-12 0.7 https://i.ytimg.com/vi/lxQokWFy5Qc/hqdefault.jpg https://i.ytimg.com/vi/lxQokWFy5Qc/hqdefault.jpg Python and Data Foundations for ML: NumPy, Pandas, EDA and Data Leakage Most machine learning bugs hide in the data code, not the algorithm. In this episode you learn the Python, NumPy, Pandas and plotting skills every ML project relies on, and how to avoid data leakage. https://www.youtube.com/embed/lxQokWFy5Qc 1013 2026-10-12T10:50:27.000Z 13 yes machine learning python for machine learning numpy tutorial pandas tutorial broadcasting numpy data leakage train test split exploratory data analysis matplotlib seaborn jupyter notebook scikit-learn missing values outliers feature types machine learning course data science foundations https://easy-explain-app.vercel.app/lesson/bbq41oFCMUw 2026-10-12 0.7 https://i.ytimg.com/vi/bbq41oFCMUw/hqdefault.jpg https://i.ytimg.com/vi/bbq41oFCMUw/hqdefault.jpg Loss Functions and Gradient Descent Explained Visually, With Numbers How does a machine actually learn? It rolls downhill. In this video we build loss functions and gradient descent from scratch, with every number worked out by hand. https://www.youtube.com/embed/bbq41oFCMUw 1044 2026-10-12T10:50:10.000Z 126 yes gradient descent loss function machine learning MSE vs MAE learning rate stochastic gradient descent mini-batch momentum Adam optimizer convex optimization saddle point local minimum NumPy gradient descent machine learning course optimization for machine learning cost function https://easy-explain-app.vercel.app/lesson/0eev3TzfMMU 2026-10-12 0.7 https://i.ytimg.com/vi/0eev3TzfMMU/hqdefault.jpg https://i.ytimg.com/vi/0eev3TzfMMU/hqdefault.jpg Maximum Likelihood, Entropy and Cross-Entropy Explained From Scratch Why does almost every classifier in machine learning minimize cross-entropy? Because it is secretly maximum likelihood in disguise. https://www.youtube.com/embed/0eev3TzfMMU 951 2026-10-12T10:49:52.000Z 69 yes maximum likelihood maximum likelihood estimation MLE entropy cross-entropy cross entropy loss KL divergence log-likelihood machine learning course machine learning math information theory MAP estimation numerical stability underflow probability for machine learning negative log likelihood https://easy-explain-app.vercel.app/lesson/IpjebaPURYo 2026-10-12 0.7 https://i.ytimg.com/vi/IpjebaPURYo/hqdefault.jpg https://i.ytimg.com/vi/IpjebaPURYo/hqdefault.jpg Regularization and Feature Engineering: Ridge, Lasso and Leak-Free Pipelines Why do models with more features often get worse on new data, and how do you fix it? In this episode we build the full toolkit: polynomial features, categorical encoding, scaling, multicollinearity, ridge regression with a full derivation of its closed form, lasso, elastic net, and the geometry that explains why L1 creates sparsity. https://www.youtube.com/embed/IpjebaPURYo 1134 2026-10-12T10:49:36.000Z 17 yes regularization feature engineering ridge regression lasso regression elastic net L1 vs L2 multicollinearity polynomial features one-hot encoding standardization feature selection scikit-learn pipeline data leakage machine learning course overfitting normal equation https://easy-explain-app.vercel.app/lesson/w8eFAvJn69c 2026-10-11 0.7 https://i.ytimg.com/vi/w8eFAvJn69c/hqdefault.jpg https://i.ytimg.com/vi/w8eFAvJn69c/hqdefault.jpg Probability Basics and Bayes' Theorem: Why a 90% Accurate Test Can Be Wrong A positive result on a 90 percent accurate medical test does not mean you are 90 percent likely to be sick. In this episode we find out why, using probability and Bayes' theorem. https://www.youtube.com/embed/w8eFAvJn69c 353 2026-10-11T14:30:29.000Z 13 yes bayes theorem probability basics conditional probability machine learning math statistics for machine learning variance and standard deviation covariance and correlation correlation vs causation base rate fallacy medical test probability bayesian updating independence probability mean median mode machine learning course python simulation probability manim https://easy-explain-app.vercel.app/lesson/molE_I8OPeY 2026-10-11 0.7 https://i.ytimg.com/vi/molE_I8OPeY/hqdefault.jpg https://i.ytimg.com/vi/molE_I8OPeY/hqdefault.jpg Partial Derivatives, Gradients, Jacobians and Hessians Explained Visually What happens when a function has many inputs? Machine learning models have millions of them, and calculus still works. https://www.youtube.com/embed/molE_I8OPeY 384 2026-10-11T14:21:08.000Z 70 yes partial derivatives gradient jacobian hessian multivariable calculus machine learning math calculus for machine learning directional derivative computational graph chain rule backpropagation basics finite differences contour plot positive definite matrix machine learning course manim https://easy-explain-app.vercel.app/lesson/IvwdUuBQQGo 2026-10-11 0.7 https://i.ytimg.com/vi/IvwdUuBQQGo/hqdefault.jpg https://i.ytimg.com/vi/IvwdUuBQQGo/hqdefault.jpg Derivatives From First Principles: Slopes, Rules and Why ML Needs Them What does it mean for a curve to have a slope at a single point? In this episode we build the derivative from scratch, using nothing but a moving secant line and a limit. https://www.youtube.com/embed/IvwdUuBQQGo 314 2026-10-11T14:08:53.000Z 88 yes derivatives calculus for machine learning derivative from first principles limit definition of derivative power rule product rule chain rule quotient rule derivative of e^x derivative of ln x numerical derivative python sympy integration area under curve math for machine learning machine learning course gradient descent intuition secant and tangent line calculus visualization https://easy-explain-app.vercel.app/lesson/-AbdWaQPATs 2026-10-11 0.7 https://i.ytimg.com/vi/-AbdWaQPATs/hqdefault.jpg https://i.ytimg.com/vi/-AbdWaQPATs/hqdefault.jpg Eigenvectors, SVD and Positive-Definite Matrices Explained Visually Most vectors get knocked off course by a matrix. A few special ones don't. Learn why eigenvectors matter, how to find them by hand, and how SVD can compress an image. https://www.youtube.com/embed/-AbdWaQPATs 403 2026-10-11T13:58:04.000Z 188 yes eigenvectors eigenvalues singular value decomposition SVD positive definite matrix linear algebra for machine learning machine learning course characteristic equation low rank approximation image compression SVD PCA NumPy linalg 3Blue1Brown style math for ML machine learning from zero to advanced https://easy-explain-app.vercel.app/lesson/FHu18ys-mhA 2026-10-11 0.7 https://i.ytimg.com/vi/FHu18ys-mhA/hqdefault.jpg https://i.ytimg.com/vi/FHu18ys-mhA/hqdefault.jpg Matrices as Transformations: See Linear Algebra Move What if a matrix isn't a table of numbers, but a machine that bends space itself? https://www.youtube.com/embed/FHu18ys-mhA 347 2026-10-11T13:40:43.000Z 95 yes matrices linear algebra matrix transformations machine learning math matrix multiplication determinant inverse matrix rank of a matrix span and basis linear independence numpy tutorial matrix vector multiplication linear algebra for machine learning 3blue1brown style machine learning course https://easy-explain-app.vercel.app/lesson/Ax-Uu4PxPbA 2026-10-11 0.7 https://i.ytimg.com/vi/Ax-Uu4PxPbA/hqdefault.jpg https://i.ytimg.com/vi/Ax-Uu4PxPbA/hqdefault.jpg Vectors and Dot Products: The Geometry Behind Every ML Prediction Every machine learning prediction starts with one tiny operation: the dot product. In this episode we build vectors from scratch and see them as arrows and as rows of data. https://www.youtube.com/embed/Ax-Uu4PxPbA 243 2026-10-11T13:31:11.000Z 58 yes vectors dot product linear algebra for machine learning machine learning course cosine similarity euclidean distance vector norm orthogonal vectors vector projection numpy dot product feature vector tensors machine learning math manim linear algebra beginners https://easy-explain-app.vercel.app/lesson/yMwfees6Z6Y 2026-10-11 0.7 https://i.ytimg.com/vi/yMwfees6Z6Y/hqdefault.jpg https://i.ytimg.com/vi/yMwfees6Z6Y/hqdefault.jpg Algebra Essentials for Machine Learning: Functions, Exponents, Logs and Sigma Every machine learning formula is built from a few algebra ideas. In this episode you learn them from zero and then read a real loss formula symbol by symbol. https://www.youtube.com/embed/yMwfees6Z6Y 399 2026-10-11T13:25:38.000Z 131 yes machine learning machine learning math algebra for machine learning functions exponents logarithms sigma notation summation notation mean squared error slope intercept form exponent laws product notation math for AI ML for beginners loss function manim https://easy-explain-app.vercel.app/lesson/aXu7_PC4vgE 2026-10-11 0.7 https://i.ytimg.com/vi/aXu7_PC4vgE/hqdefault.jpg https://i.ytimg.com/vi/aXu7_PC4vgE/hqdefault.jpg What Is Machine Learning? A Visual Introduction for Beginners How does your email know which messages are spam without anyone writing every rule? This video is a visual, math-light introduction to machine learning. https://www.youtube.com/embed/aXu7_PC4vgE 332 2026-10-11T13:16:36.000Z 95 yes machine learning what is machine learning machine learning for beginners ml introduction AI vs machine learning deep learning data science supervised learning unsupervised learning reinforcement learning self-supervised learning machine learning course ml workflow visual explanation python machine learning loss function https://easy-explain-app.vercel.app/lesson/y8Zrg_3_p7A 2026-10-11 0.7 https://i.ytimg.com/vi/y8Zrg_3_p7A/hqdefault.jpg https://i.ytimg.com/vi/y8Zrg_3_p7A/hqdefault.jpg Why the Pythagorean Theorem Actually Works (Visual Proof) You've memorized a squared plus b squared equals c squared, but do you know why it's true? https://www.youtube.com/embed/y8Zrg_3_p7A 51 2026-10-11T08:48:22.000Z 1735 yes pythagorean theorem why pythagorean theorem works pythagorean theorem proof visual proof math shorts geometry right triangle a squared plus b squared rearrangement proof math explained proof without words learn math mathematics 3blue1brown style shorts https://easy-explain-app.vercel.app/lesson/iB_zMCG3CMM 2026-10-11 0.7 https://i.ytimg.com/vi/iB_zMCG3CMM/hqdefault.jpg https://i.ytimg.com/vi/iB_zMCG3CMM/hqdefault.jpg Correlation vs Covariance: What's the Difference? Do two things move together? Covariance and correlation answer that, but only one is easy to compare. https://www.youtube.com/embed/iB_zMCG3CMM 63 2026-10-11T08:45:31.000Z 929 yes correlation covariance statistics data science correlation vs covariance pearson correlation machine learning math explained manim statistics for beginners scatter plot portfolio diversification feature selection correlation is not causation shorts