ai-hands-on

Why we included this project

Learning AI from scratch usually means wading through theory and then hoping the pieces click. This collection of notebooks takes the opposite route: you do the work. It starts with the math that makes machine learning work, moves through PyTorch tensors and hand-built networks, then into attention, transformers, RAG pipelines, and OCR. Because the notebooks are meant to be run and modified, you get a feel for the internals, not just the API calls. Beginners get a clear path to follow, and engineers who already work with models can jump straight to the parts they're missing, like building a RAG pipeline end to end or writing normalization and optimizers from scratch. If you want to know how the systems you rely on are actually assembled, this is a good place to spend an afternoon.