// the find
DataTalksClub/machine-learning-zoomcamp
Learn ML engineering for free in 4 months! Register here 👇🏼
A free, structured 4-month course teaching ML engineering from scratch — linear regression through Kubernetes deployment. Targets software engineers who can code but haven't touched ML yet. The repo is the course materials themselves: Jupyter notebooks, markdown lessons, homework, and working code samples.
The deployment coverage is unusually good for a free course — you get Docker, FastAPI, AWS Lambda, Kubernetes, and TensorFlow Serving all with actual working code, not just slides. The progression from scratch implementations (linear regression by hand) to scikit-learn to production deployment is well-sequenced, not just a random topic dump. Workshop directories ship with complete, runnable projects including Dockerfiles and lock files, so the examples actually work. The community infrastructure (Slack, annual cohort, peer review) is real and active — 13k stars and 3k forks aren't just GitHub vanity.
The deep learning module is shallow relative to what the deployment modules promise — you get intro CNNs and transfer learning, but nothing on LLMs, fine-tuning, or modern inference patterns that are actually in demand right now. The course leans heavily on TensorFlow/Keras and only added PyTorch as an afterthought; the PyTorch content is thin. Module 7 is missing entirely from the syllabus and directory tree, which is a disorienting gap. Self-paced learners get no certificate and no graded feedback, so there's nothing to distinguish your completion to an employer beyond a GitHub project you built yourself.