Learn AI from scratch — a path that adapts to you

There is no «AI course for everyone». Someone who can code but not do maths needs a different path from someone who knows maths but has never written code. That is why AI-grafen starts with a diagnostic instead of a syllabus.

The path runs through knowledge nodes with prerequisites: algorithmic thinking → Python → statistics and probability → vectors and matrices → gradient descent → neural networks → backpropagation → PyTorch → attention → transformers → language models.

Every node has explanations at several depths, exercises that measure mastery and a tutor that helps you get there yourself. Labs run real code — in the browser or in a sandbox.

Already a programmer? Go straight to AI for developers, RAG, agents or evals. How we know what you know.