GitHub radar
Learn generative AI basics for free: Microsoft's 21-lesson course
A free 21-lesson course from Microsoft explains how large language models work and how to build apps with them — after it you'll understand what your agent is actually doing.
A free course with 21 lessons, each covering one generative AI topic — from prompts to meaning-based search (RAG) and running your own models.
When it helps
Useful if you use an AI agent daily but don't understand how prompts, RAG, and vector databases actually work. It won't teach you to code or give you a finished app.
Pros
- 21 topics, each self-contained — start at any lesson
- Code examples in both Python and TypeScript
- Can run fully offline via Foundry Local, no paid API key needed
- Free, from Microsoft, with a video for each lesson
Cons
- Code examples target Azure OpenAI and OpenAI API, not Claude Code
- Doesn't teach programming — only explains concepts
- Requires a GitHub account to fork the repo
How to set it up — step by step
- 1Go to https://github.com/microsoft/generative-ai-for-beginners and star it so you don't lose it.
- 2Fork the repository to your own GitHub account.
- 3Send your agent the text from the block below.
- 4Check that the agent opened and explained lesson 01 in plain words.
- 5Run the lesson's code example through Foundry Local — no paid API key needed.
- 6Move on to whichever lessons interest you, in any order.
Text for your agent
Copy this and send it to your agent — Claude Code, Codex, any of them:
Set up the course from https://github.com/microsoft/generative-ai-for-beginners — open Lesson 01 (Introduction to Generative AI and LLMs) with me, explain it in your own words, and run the code example through Foundry Local so we don't need a paid API key.
The lessons themselves don't need a powerful computer — the code calls an API, there's no heavy computation on your machine. If you pick the offline option (Foundry Local), a regular laptop is enough for the small models used in the examples.
Open on GitHub▌ More finds