Ai Engineering From Scratch

Free

Learn it. Build it. Ship it for others.

FreeFree tier

Type

Open Source

Visit Website GitHub (25.5k stars)

LinksX

About Ai Engineering From Scratch

AI Engineering from Scratch is a free, open-source curriculum (MIT license) that builds every core AI algorithm from the ground up. It consists of 503 lessons across 20 phases, covering topics from linear algebra to autonomous agents, in Python, TypeScript, Rust, and Julia. Each algorithm is constructed from raw math before any framework is introduced, ensuring deep understanding of underlying principles. Lessons are authored with AI assistance and human-reviewed, with citations to primary sources (RFCs, specs, research papers). The curriculum runs entirely on your own machine with no paywall or signup required. The website is deliberately plain (hand-written HTML, CSS, and vanilla JavaScript) and hosted on Vercel. As of now, the GitHub repository has over 25,457 stars.

Key Features

Free and open source (MIT license)

503 lessons across 20 phases

Build every algorithm from raw math before frameworks

Available in Python, TypeScript, Rust, and Julia

Each lesson produces a reusable artifact (prompt, skill, agent, MCP server)

Human-reviewed with citations to primary sources

Runs entirely on your own machine with no paywall or signup

Plain HTML/CSS/JS site with hand-written code

Pros & Cons

Pros

Comprehensive 503-lesson curriculum covering full AI spectrum
Free and open source with no gated content or upsells
Builds understanding from raw math, not just API usage
Supports four programming languages
Human-reviewed lessons with citations to original sources
Every lesson yields a practical artifact you keep
Runs locally on your own machine with full control

Cons

No video lessons – text and code only
Requires strong self-discipline and independent problem-solving
May be challenging for absolute beginners without math background
Deliberately plain website design with no visual aids
No interactive exercises or graded assignments

Best For

Self-paced learning of AI engineering fundamentalsDeep understanding of algorithm internals via math-first approachBuilding a portfolio of reusable AI artifactsTransitioning from framework-dependent knowledge to foundational skillsOpen-source contribution and curriculum improvement

FAQ

Is AI Engineering from Scratch free?

Yes, it is completely free and open source under the MIT license. There is no paywall, no signup, and no course upsells.

What languages are used in the curriculum?

Lessons are available in Python, TypeScript, Rust, and Julia, depending on what fits the concept best.

How many lessons are there?

There are 503 lessons organized into 20 phases, covering from linear algebra to autonomous agents.

Do I need to install anything?

Yes, the curriculum runs on your own machine. You clone the GitHub repository and run the code locally.

How are lessons structured?

Each lesson follows a loop: read the problem, derive the math, write the code, run the test, and keep the artifact. Lessons are text-based with runnable code.

Is the curriculum reviewed?

Lessons are authored with AI assistance and human-reviewed against primary sources such as RFCs, specs, or research papers. Corrections are tracked openly on GitHub.