harvard-edge/cs249r_book
OPEN TEXTBOOK ON ML SYSTEMS
Open textbook series on machine learning systems (MLSys) with practical tools. Teaches AI engineering: building efficient, reliable, and safe AI systems in real-world conditions.
// KEY FEATURES
- Four volumes: foundations, scaling, agentic AI, physical AI
- TinyTorch: students build ML machinery themselves
- MLSys·im simulator and hardware kits for experiments
- Focus on constraints: memory, latency, energy, fault tolerance
// DEEP DIVE: HOW IT WORKS
Repository combines four volumes of open textbook on ML systems and practical tools. Material built on principle that constraints drive architecture: each section connects ML concepts to system implementation — from distributed training and quantization to serving, monitoring, and energy consumption. Started as Harvard CS249r student notes in 2023.
Theory complemented by executable components: TinyTorch makes students implement framework core themselves, MLSys·im simulates systems from microcontrollers to datacenters, hardware kits enable testing on real devices. Volume I accepted for print by MIT Press, volumes II–IV in Preview/Dev. Teaches AI engineering through first principles and recomputable numbers.
Oriented to education, not production: no ready libraries, CLI, or API. Requires time for course and hardware for kits/simulator. Translations lag behind English version.
// USE CASES
Best for: Students, engineers, and ML systems educators learning AI engineering from scratch
- Learn AI engineering from first principles through structured course with practice
- Understand how memory, latency, and energy constraints define ML system architecture
- Prepare ML systems teaching material with ready labs, simulator, and kits
New repositories every 30 minutes
REDDYX AI scans GitHub 24/7 and ships the best AI/ML/Web3 projects to Telegram.
Join on Telegram