Harvard's Free ML Systems Curriculum Just Expanded to Four Volumes
Prof. Vijay Janapa Reddi’s Machine Learning Systems project has grown into a full curriculum for understanding what happens when an ML model has to become an actual system. The official site now describes it as a complete curriculum for AI engineering.
Volume I goes from ML system fundamentals and training to model optimization, hardware acceleration, benchmarking, deployment, serving and operations. Volume II moves into the harder systems layer: compute infrastructure, distributed training, communication, reliability, inference at scale and fleet operations.
And it isn’t just two textbooks.
🔬 34 interactive labs that run in the browser
⚙️ TinyTorch for building the machinery from tensors upward
🪖 hardware kits for hands-on deployment
🏗️ MLSys·im for reasoning about systems and infrastructure
🧪 StaffML for ML systems design practice
The GitHub project is already sitting at ~28K stars.
A lot of people learn how to train a model. Far fewer learn what it takes to make that model work reliably as part of a real system.
This is real, from a legitimate and highly credentialed source — but there's one structural claim worth updating, and a couple of specific numbers I couldn't independently pin down.
Confirmed: Prof. Vijay Janapa Reddi is a real Harvard professor with exactly the credentials that make this worth trusting — Associate Professor at Harvard, Inference Co-chair for MLPerf, and a founding member of MLCommons, with a Harvard PhD and a substantial research/awards record in computer architecture and edge/mobile ML systems.
The project and its five-part ecosystem — all confirmed exactly. The GitHub repo (harvard-edge/cs249r_book) itself frames it exactly as your post does: the textbook, TinyTorch (build-from-scratch framework, confirmed literally as "from tensors to systems"), hardware kits, MLSys·im ("a modeling engine for infrastructure you can't physically access" — confirmed verbatim), and StaffML ("physics-grounded systems design questions across cloud, edge, mobile, and TinyML" — confirmed verbatim). Every named component in your post is real and accurately described.
One structural claim worth updating: this is no longer a two-volume project. The repository's current title is "Machine Learning Systems: Foundations, Scaling, Agentic AI, and Physical AI (Vols I–IV)" — meaning the project has expanded to four planned volumes, not two. What your post calls Volume I and Volume II are accurate as far as they go (Volume I is content-complete/in editorial polish; Volume II, "At Scale," is under active development) — but Volumes III (Agentic AI) and IV (Physical AI) exist as announced scope beyond what's described here. Not wrong, just understating how far the project's ambitions now extend.
"34 interactive labs" — I couldn't independently confirm this exact count. The labs are real (interactive Marimo notebooks that run in-browser, letting you change a parameter and see what breaks — genuinely matching "run in the browser"), but I didn't find a source confirming the specific number 34.
"~28K GitHub stars" — plausible but unconfirmed at that exact figure. The one hard data point I found states the project hit its 10,000-star goal by December 2025. Growing to ~28K by September 2026 would be a roughly 3x increase in under a year — aggressive, but not implausible for a project this actively promoted (MIT Press deal, university adoption push). Worth verifying against the live repo before repeating the exact number.
One extra fact worth adding, since the post's "$2,000 course" framing is about value: the textbook content is explicitly licensed CC BY-NC-SA 4.0 — free and open for non-commercial use — while a formal MIT Press print edition of Volume I is separately planned for 2026. So the free/open framing holds for the online curriculum regardless of the commercial print book.