Open Source · MIT License

AI Engineering
from Scratch

230+ hands-on lessons across 20 phases. From linear algebra to autonomous agent swarms. Python, TypeScript, Rust, Julia. Every lesson produces something reusable.

-- Lessons
-- Phases
4 Languages
-- Complete

Why This Course

Full Scope

Math, ML, deep learning, NLP, vision, speech, transformers, LLMs, agents, swarms. Everything in one place.

Multi-Language

Python, TypeScript, Rust, Julia. Build in whatever language fits the problem, not just one.

Build, Don't Watch

Every lesson produces reusable prompts, skills, agents, or MCP servers. A portfolio, not a certificate.

Scratch First

Implement from first principles, then learn the framework. Understand why, not just how.

The Journey

20 phases. Each builds on the last. Click any phase to explore its lessons.

Progress Roadmap

Track overall course completion.

How Each Lesson Works

01

Motto

One-line core idea that frames the lesson.

02

Problem

Why this matters. What problem does it solve.

03

Concept

Visual diagrams and intuition-first explanations.

04

Build It

Implement from scratch. No frameworks yet.

05

Use It

Same thing with real frameworks and tools.

06

Ship It

The prompt, skill, or agent this lesson produces.

Glossary

Key AI terms, demystified. What people say vs. what they actually mean.

Start Learning

git clone https://github.com/rohitg00/ai-engineering-from-scratch.git

Prerequisites: You can write code (Python or any language) and you want to understand how AI actually works.