Tap a word, watch the model rank it
no code · fully interactive · the whole stack
Understand how modern LLMs actually work, by building intuition, not watching videos.
One connected arc, from “a word is a vector” through attention, transformers, scaling, RAG and agents, down to the GPUs underneath. Every lesson opens with a wall you can feel, then hands you the tool that breaks it.
Pick a path
Three routes through the course, by what you want out of it. Jump in anywhere, they overlap.
Core Mechanics
How LLMs actually work, end to end, tokens and vectors to attention, scaling, RAG, and the limits.
35 lessons · ~2.9 h
0/35 done
Start →Daily AI Use
Drive AI tools well, context, agents, permissions, and habits that get more from every prompt. No code.
23 lessons · ~1.9 h
0/23 done
Start →Production Builder
Ship and run LLM features, the pipeline, observability, evals, data, security, and trustworthy UX.
18 lessons · ~1.5 h
0/18 done
Start →The journey
0 of 63 core · 0%The Big Picture
See it work firstRepresentation
A word is a vectorPrediction & Learning
How it gets goodArchitecture
Attention & the transformer blockOperating & Scaling
Decoding, context, scaling lawsA System in the World
Frozen model → RAG → agentsAgents in Practice
Drive the tools wellLLMs in Production
Ship, watch, evaluate, improvePhysical Substrate
GPUs & datacentersElective Rooms
Off the main arcelectivesYou understand how AI works, end to end.