Learn AI engineering by playing with it

From “computers only understand numbers” to shipping production AI systems — through interactive visualizations, free Colab notebooks, and a real Gemini API key. Zero AI background assumed.

Start Learning Browse the Course

modules: 13
lessons: 34
interactive playgrounds: 18
forever: $0

Why this course exists

Most AI tutorials assume you already know the jargon. This one assumes you know nothing — and shows you everything.

01
Why before how
Every lesson starts with the problem a technology solves. You'll never memorize a definition without knowing why it exists.
02
Play, don't just read
Tokenizers, embedding graphs, attention maps, RAG pipelines — everything is a playground you can poke at until it clicks.
03
First principles, zero assumptions
Built for QA engineers, backend devs, support engineers, and students. If you can read code, you can learn AI engineering here.

The roadmap

13 modules, each building on the last — from raw bits to systems users rely on. No floor skipped, no magic allowed.

Start — zero AI knowledge Finish — you ship production AI

092 lessonsModel Context ProtocolA universal port for connecting AI to tools 082 lessonsAgentsModels that plan, use tools, and act 074 lessonsRetrieval-Augmented GenerationGiving models knowledge they were never trained on 063 lessonsThe PlaygroundBuild an OpenAI-style playground 056 lessonsLarge Language ModelsAttention, sampling, and next-token prediction

102 lessonsVector DatabasesSearching by meaning at scale 112 lessonsFine-TuningAdapting models with LoRA and friends 120 lessonsProduction AIShipping AI systems that survive real users

002 lessonsIntroductionWhat AI Engineering is and why it matters 013 lessonsComputers Understand NumbersBinary, bytes, and how text becomes numbers 023 lessonsTokenizationHow language models read text 032 lessonsAutoencodersHow machines learn compact representations 043 lessonsEmbeddingsTurning meaning into geometry 056 lessonsLarge Language ModelsAttention, sampling, and next-token prediction 063 lessonsThe PlaygroundBuild an OpenAI-style playground 074 lessonsRetrieval-Augmented GenerationGiving models knowledge they were never trained on 082 lessonsAgentsModels that plan, use tools, and act 092 lessonsModel Context ProtocolA universal port for connecting AI to tools 102 lessonsVector DatabasesSearching by meaning at scale 112 lessonsFine-TuningAdapting models with LoRA and friends 120 lessonsProduction AIShipping AI systems that survive real users

Swipe to explore all 13 modules →

Real tools, not toy examples

Every code snippet in the course actually runs. You'll finish having used the same products working AI engineers use daily.

Interactive lessons

18 hand-built playgrounds run right in your browser — no install, no account.

Colab notebooks

Hands-on modules open a real notebook in one click. Run every cell, break things safely.

Free Gemini API key

All code examples run against Gemini's free tier — real API calls, zero spend.

Going deeper

Advanced essays for engineers past the basics — inference internals, eval strategy, and production war stories.

All posts

We'll put interesting stuff here.

Ready to stop treating AI as magic?

Start with a single question — how does a computer read the letter A? — and end shipping a production RAG system.

Start Module 0