Learn RAG and AI Agents by Actually Building Them

RAG and agents are the two AI skills companies are overwhelmingly hiring for right now. No cap: recruiters ctrl+F for these. This free project teaches you both the practical way, by building a working AI agent you can extend, ship, and flex in an interview.

Build a real AI agent in TypeScript that lives in your repo, not in a chat window
Implement RAG from scratch so you can explain retrieval like you built it. Because you did.
Start simple, then extend it as far as you want -- the code is yours forever

Open any AI engineering job post.
You'll see the same two skills: RAG and agents.

Chatting with an LLM stopped being impressive years ago. Everyone can do that. The developers getting hired can build retrieval pipelines that ground answers in real data, and agents that reliably do useful work. Most developers have never built either. This project walks you through both, in code, on a system you build yourself.

What You'll Build

One project, built up piece by piece. Each step adds a skill that shows up in real AI engineering work.

Step 01

From Chat User to AI Developer

Make your first programmatic LLM call with the Gemini API and cross the line from using AI tools to building with them in code.

Step 02

Take Control of the Agent

Use system prompts and structure to turn a generic chatbot into a focused AI advisor with a clear role, voice, and behavior.

Step 03

Give It Memory with RAG

Implement a simple retrieval workflow so your agent answers from real source material instead of confidently making things up.

Step 04

Ground It in Real Data

Feed in YouTube transcripts as context so your agent gets genuinely useful on a real topic instead of just sounding confident.

Step 05

See What Real AI Products Look Like

Connect what you built to production-style systems, see how these ideas scale, and get a clear map of what to learn next if you want to go deeper.

What you'll walk away with

A working system and the two skills the market is paying for. That's the list, and it's a good list.

A working AI agent you built yourself. Simple on purpose, and built to be extended.
RAG, implemented by hand -- retrieval, context, and grounded answers you can explain in an interview
Agent fundamentals -- system prompts, roles, and structure that make an LLM actually useful
Code you keep -- extend it into a portfolio project or steal the patterns for work

Learn the two skills every AI job post asks for

RAG and agents aren't going anywhere. They're the foundation of nearly every AI product being built right now, and this is the lowest-effort way to stop nodding along in meetings and start actually knowing how they work. TypeScript, the Gemini API (free tier), and zero black-box magic. Free, and yours to keep.

Instant access: Start right away

Keep the code: Everything you build is yours

No risk: Free to join, unsubscribe anytime