AI Engineering from Scratch
A free, MIT-licensed course that teaches AI engineering by building models, retrieval, and agents from scratch
- Stars
- 76.3k
- Forks
- 13.5k
- Licence
- MIT
- Language
- Python
- Last update
- 4 days ago
- Type
- Learning resource
What it is
This is a structured course with 542 lessons across 20 phases, covering math foundations, machine learning, LLM engineering, and agent engineering. Lessons are taught on a companion website, with a coding agent, or by running the local code in the repository. The README says the goal is to implement model internals, retrieval pipelines, and agent runtimes, test them, and keep the code and evaluation results.
- Work through the setup and tooling lesson, then follow a learning path
- Implement a retrieval evaluation project and run its staged tests with the provided script
- Compare gradient descent, momentum, and Adam using the included optimizer code
- Build the agent loop lessons into a small agent runtime
Great for
- Students who want to understand how LLMs, retrieval, and agents work internally
- Developers who prefer learning by writing code from scratch
- Learners following the listed paths: model foundations, LLM systems, or agents and delivery
- People who want hands-on projects with local graders to check their work
Not ideal for
- People who only want a quick ready-made AI tool; this is a course, not an application
- Complete beginners to programming, since the README lists prerequisites only in part and the study guide section was not included in the excerpt
- Anyone expecting a single installable package; the README points to lessons and project folders rather than one install command
How to get it
- 1
Clone and build from source
You'll need: Python 3.10+
Clone the repository
git clone https://github.com/rohitg00/ai-engineering-from-scratch - 2
Run it
- Clone the repository and read the setup and tooling lesson on the website
- Choose a learning path from the README table
- Initialize a project with python3 scripts/project_test.py retrieval-evaluation-lab --init learning-artifacts/retrieval-evaluation-lab
- Implement stage 1 and run the stage test with --stage 1 --strict
- 3
Get the most out of it
- Use the interactive gradient descent lab on the website to adjust the learning rate
- Compare optimizer implementations in phases/01-math-foundations/08-optimization/code/optimizers.py
- Follow the project challenges and keep the per-query comparison output as evidence of your work
- Use the cheatsheet and PDF assets in the latest release as quick references
Downloads
Strengths
- Free, open source under MIT, with lessons available on the website and locally
- Projects include staged starters, reference solutions, and local graders for self-checking
- Covers Python, TypeScript, Rust, and Julia according to the README
Trade-offs
- The README excerpt is truncated, so the full curriculum, prerequisites, and setup steps are not visible here
- Projects start failing by design until the learner implements each stage, which can be frustrating without guidance
- The course is large (542 lessons), so the time commitment is substantial
How it compares
| Project | Stars | Type | Level | Updated | Best for |
|---|---|---|---|---|---|
| ai-engineering-from-scratch (this) | 76.3k | Learning resource | intermediate | 4 days ago | Students who want to understand how LLMs, retrieval, and agents work internally |
| Prompt-Engineering-Guide | 79.8k | Learning resource | beginner | 8 months ago | People who are new to prompting and want structured lessons with examples |
| awesome-ai-engineering | 22.5k | Curated list | beginner | 15 days ago | Developers who want an overview of the AI engineering topics covered in the description |
| AI-For-Beginners | 70.5k | Learning resource | beginner | 13 days ago | Beginners who want a structured, self-paced introduction to AI |
| ai-engineering-hub | 40.8k | Learning resource | intermediate | 5 days ago | Developers who want working examples of RAG or agent systems to adapt |
| fashion-mnist | 12.9k | Learning resource | beginner | 4.4 years ago | Machine learning students comparing classifiers on a familiar format |
Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.