// the find
chiphuyen/machine-learning-systems-design
A booklet on machine learning systems design with exercises. NOT the repo for the book "Designing Machine Learning Systems", which is `dmls-book`
A 2019 booklet by Chip Huyen covering the four stages of ML system design: project setup, data pipeline, modeling, and serving. The author has since written a full O'Reilly book on the same topic and openly says this is superseded — it's now mainly useful as interview prep for its 27 open-ended ML systems design questions.
The 27 interview questions are genuinely good and cover the kind of system-level thinking that coding interviews miss — trade-offs in data collection, latency vs. accuracy, monitoring drift. Community answers in the repo give multiple perspectives on each question. The research-vs-production framing cuts through a lot of confusion that trips up people moving from notebooks to deployed models. Chip Huyen's writing is clear and direct, no filler.
The author's own README says to read the newer book instead — this is archived material from 2019 and the ML tooling landscape has changed significantly since then. Last pushed in 2023, and that was almost certainly just metadata, not content updates. The community answers vary wildly in quality with no curation. If you want this content in a usable form, you're better off going to the published HTML at huyenchip.com than cloning the repo.