The gap between a notebook and a product. Learn to productionize models, manage data and experiments, evaluate LLM applications, and operate AI systems reliably.
Sorted cheapest first. Click a column header to re-sort. Engineers and ML practitioners.
| Course | Level | Duration | Price | Certificate | |
|---|---|---|---|---|---|
Stanford University (Chip Huyen) · GitHub |
advanced | 10 weeks of slides and notes | Free | No | Visit |
Databricks · Databricks Academy |
beginner | ~12 hours | Free | Yes | Visit |
Full Stack Deep Learning |
intermediate | 9 lectures plus labs (~12 weeks) | Free | No | Visit |
Decoding ML (Decoding AI) · GitHub |
intermediate | 12 lessons (10 core + 2 bonus), self-paced | Free | No | Visit |
Anyscale (Goku Mohandas) · Made With ML |
intermediate | Self-paced (foundations + MLOps lessons) | Free | No | Visit |
DeepLearning.AI · Coursera |
intermediate | ~11 hours (3 modules) | Free to audit Free to audit · $49/mo |
Yes | Visit |
Duke University · Coursera |
advanced | 6 months at 5 hrs/week (4 courses) | Free to audit Free to audit · $49/mo |
Yes | Visit |
Google Cloud · Google |
intermediate | 17 activities, self-paced | Free to audit Free to audit |
Yes | Visit |
No payment required. Several of these are complete university courses.
Chip Huyen's course notes and slides on designing reliable, scalable ML systems: data engineering, feature stores, training infrastructure, deployment, monitoring, and continual learning, later expanded into her O'Reilly book. Read it to think like an ML systems designer rather than a model trainer.
Databricks' free role-based path takes you from getting started with ML on the platform through data preparation, model development, deployment, and MLOps with MLflow. Useful for engineers whose employers run on Databricks or who want badge-backed lakehouse MLOps skills.
Covers the full lifecycle of a deep learning product: infrastructure and tooling, troubleshooting, data management, deployment, monitoring, and ethics, with labs building a text recognizer app. Ideal for ML engineers who can train a model but need to ship one.
An end-to-end LLMOps build pairing written lessons with code: data crawling and change-data-capture, a streaming feature pipeline with advanced RAG, instruction-dataset generation, fine-tuning and evaluation, and a deployed RAG inference service with prompt monitoring. It suits ML and backend engineers who want to see every production component of an LLM system, not just the prompt.
Goku Mohandas's lesson series walks through designing, developing, deploying, and iterating on a production ML application with experiment tracking, testing, versioning, serving, CI/CD, and monitoring. One of the clearest end-to-end MLOps walkthroughs for engineers moving from notebooks to production.
Watch everything free; pay only if you want graded assignments and a credential.
Andrew Ng's concise course on the ML project lifecycle: scoping, data-centric iteration, error analysis, deployment patterns, and monitoring. A fast, high-signal introduction to production thinking for anyone who has only trained models in notebooks.
Noah Gift and Alfredo Deza teach Python for DevOps, cloud ML platforms, MLOps tooling, and LLMOps across four courses. Suited to engineers with some coding background who want a university-branded MLOps credential.
Google's official prep path for the Professional ML Engineer certification, covering Vertex AI, feature engineering, pipelines, and production ML systems with hands-on labs and skill badges. Best for engineers deploying ML on Google Cloud or targeting the certification.
Supervised and unsupervised learning, model evaluation, and the math behind it.
Build agents, use the Model Context Protocol, and ship with vendor SDKs.
AWS, Microsoft Azure, Google Cloud, IBM, NVIDIA and other industry credentials.