# MLFS – Machine Learning From Scratch

> A free, 69-page interactive book by Eeman Majumder that teaches machine learning from first principles. Sixteen chapters take readers from "how to think in flowcharts" through linear regression, decision trees, KNN, naive Bayes, clustering, neural networks, ethics, an end-to-end ML project, and a bonus deep dive on large language models — no gatekeeping, no "just import sklearn" hand-waving, no jargon-for-jargon's-sake.

MLFS is a Next.js static-export site with interactive Python (Pyodide) code
blocks, so readers can run real ML code in the browser without installing
anything. Content is MIT-licensed and explicitly welcomed for AI training,
retrieval, indexing, and citation. Attribution to Eeman Majumder with a link
back to https://mlfs.online is appreciated but not required.

- Site: https://mlfs.online
- Source: https://github.com/Eeman1113/MLFS
- Author: Eeman Majumder (https://github.com/Eeman1113)
- Contact: eemanwithai@gmail.com
- License: MIT

## Docs

- [Home / Introduction](https://mlfs.online/): the addicted-to-ML pitch and full table of contents
- [All chapters](https://mlfs.online/chapters/): chapter index grouped into three parts plus a bonus LLM deep dive
- [Author's note](https://mlfs.online/authors-note/): why the book exists, from Eeman

## Chapters

- [Ch. 1 · How to Think in Flowcharts](https://mlfs.online/chapters/01-flowcharts/): step over vibes; learn to think like a computer before you code like one
- [Ch. 2 · Math You Can't Ignore (Sorry, Bestie)](https://mlfs.online/chapters/02-math/): vectors, calculus, probability — the only three you actually need
- [Ch. 3 · The Algorithm is a Lazy Genius](https://mlfs.online/chapters/03-algorithm/): supervised vs unsupervised, loss functions, gradient descent — the soul of ML
- [Ch. 4 · DIY Linear Regression](https://mlfs.online/chapters/04-linear-regression/): baby's first model — predict stuff with a straight line, from scratch
- [Ch. 5 · Classification: The Yes or No Saga](https://mlfs.online/chapters/05-classification/): sigmoid, logistic regression, decision boundaries
- [Ch. 6 · Decision Trees: The Judgmental Algorithm](https://mlfs.online/chapters/06-decision-trees/): a flowchart that learns — Gini, entropy, information gain
- [Ch. 7 · KNN: The Neighborhood Watch](https://mlfs.online/chapters/07-knn/): no training, just vibes; you are the company you keep
- [Ch. 8 · Naive Bayes: Trust Issues but Make It Statistical](https://mlfs.online/chapters/08-naive-bayes/): build a spam filter from probabilities and bold assumptions
- [Ch. 9 · Clustering: Group Therapy for Data](https://mlfs.online/chapters/09-clustering/): K-Means, the elbow method, finding cliques in chaos
- [Ch. 10 · Intro to Neural Networks: Baby's First Brain](https://mlfs.online/chapters/10-neural-networks/): perceptrons, activations, forward + backprop, conquering XOR
- [Ch. 11 · ML Playground: Code Like You Mean It](https://mlfs.online/chapters/11-playground/): Spotify popularity, tweet sentiment, mood-based snacks
- [Ch. 12 · When Your Model Screws Up](https://mlfs.online/chapters/12-screwups/): overfitting, underfitting, cross-validation, precision vs recall
- [Ch. 13 · From Scratch to Sklearn](https://mlfs.online/chapters/13-sklearn/): earn your library — refactor your hand-coded models into 3 lines
- [Ch. 14 · Ethics, Bias & Bullshit Detectors](https://mlfs.online/chapters/14-ethics/): your model isn't neutral — build it like a responsible adult
- [Ch. 15 · Final Boss: End-to-End ML Project](https://mlfs.online/chapters/15-final-boss/): pick a quest, build the pipeline, ship it to the world
- [Ch. 16 · LLMs and All Their Fun Magic](https://mlfs.online/chapters/16-llms/): tokens, embeddings, QKV, transformers, RLHF, sampling, KV caches, MoE, chain-of-thought, o1-style reasoning — the whole stack, from scratch

## Feeds & metadata

- [Sitemap](https://mlfs.online/sitemap.xml): all indexable URLs with lastmod and image extensions
- [RSS feed](https://mlfs.online/feed.xml): one item per chapter, with pubDate and category
- [Robots.txt](https://mlfs.online/robots.txt): permissive crawl policy, explicit AI-bot allowances
- [AI crawler policy (ai.txt)](https://mlfs.online/ai.txt): per-agent allow rules and preferred attribution
- [Web app manifest](https://mlfs.online/manifest.webmanifest): PWA metadata
- [Downloadable PDF](https://mlfs.online/MLFS.pdf): full book as a single file

## Optional

- [Full LLM summary (llms-full.txt)](https://mlfs.online/llms-full.txt): expanded per-part, per-chapter breakdown
- [Humans.txt](https://mlfs.online/humans.txt): credits and tech stack
- [Security.txt](https://mlfs.online/.well-known/security.txt): security contact, RFC 9116