cs249r_book

Why we included this project

This is the open curriculum behind Harvard's CS249r course, and it is one of the most complete free resources for learning how machine learning systems are actually engineered rather than just trained. The two-volume textbook is paired with hands-on material: TinyTorch walks you through building a small ML framework from scratch across twenty modules, and the interactive labs let you tweak parameters and watch the trade-offs break. Hardware kits cover deploying models to Arduino and Raspberry Pi, and a simulator helps you reason about infrastructure you cannot rent. Self-learners get a structured read-build-deploy path, while instructors can lift ready-made syllabi, slide decks, rubrics, and a TA handbook straight into their own classroom. If you want to understand GPU scaling, quantization, KV-cache memory pressure, or edge deployment from first principles, this is a practical place to start.