Neural Networks: Zero to Hero

Self-directed

Andrej Karpathy’s course, worked through from scratch: micrograd, makemore and a small GPT, written without high-level libraries.

Code on GitHub