An open-source, end-to-end AI engineering curriculum that teaches how AI systems are built by having learners implement core ideas from scratch and turn each lesson into reusable artifacts like prompts, skills, agents, or MCP servers.
This repository is a structured AI engineering curriculum rather than a single library or app. It spans 20 phases and 503 lessons across Python, TypeScript, Rust, and Julia, covering topics from math foundations and machine learning through transformers, LLM engineering, agents, multi-agent systems, production, ethics, and capstone projects. The README says each lesson is designed to produce a reusable output, such as a prompt, skill, agent, or MCP server.
It addresses the common gap between using AI tools and understanding how to build them professionally. According to the README, many learners can use AI products, but far fewer feel prepared to engineer AI systems end to end; this curriculum is meant to close that gap by connecting theory, implementation, and practical outputs in one path.
Conceptually, the curriculum is organized as a layered progression. Learners start with setup, tooling, math, and machine learning basics, then move into deep learning, vision, NLP, transformers, generative AI, LLMs, multimodal systems, tools and protocols, agents, autonomous systems, multi-agent swarms, production, ethics, and final projects. Each lesson follows a repeated pattern: understand the problem, derive the underlying idea, build it from scratch, compare or apply it with standard libraries, and keep the resulting artifact.
It is gaining attention because it combines several things the current AI community values: free and open-source access, a very large lesson count, multiple languages, and an emphasis on building real artifacts rather than passively watching tutorials. The README also frames it around hot areas such as agents, MCP, LLM engineering, multimodal AI, and production systems, which aligns with current interest in practical AI engineering.
Based on the README, the closest alternatives are conventional AI tutorials, scattered blog posts, standalone courses, or reading papers and building only small demos. The repository positions itself against that fragmented approach, and it also references the creator’s separate Agent Memory project as a related open-source work, though that is not presented as the curriculum’s main alternative.
AI-explained · grounded in each repo's README