📄️ Table of contents
Free discovery course on MLOps in 6 lessons. Why most models never ship, how to make training reproducible, which deployment pattern to choose, what to monitor, and how teams organise around it.
📄️ 1. Why models never ship
Lesson 1 of the free MLOps course: the seven real reasons projects fail, why accuracy is almost never one of them, and the questions to answer before writing any code.
📄️ 2. Reproducibility
Lesson 2 of the free MLOps course: the five things that must be versioned, what experiment tracking and model registries do, and what a feature store is actually for.
📄️ 3. Deployment patterns
Lesson 3 of the free MLOps course: choosing between batch, real-time and edge serving, and releasing a model safely with shadow deployment, canary releases and A/B tests.
📄️ 4. Monitoring
Lesson 4 of the free MLOps course: the four monitoring layers, the difference between data drift and concept drift, how to cope with delayed ground truth, and when to retrain.
📄️ 5. Teams, cost, governance
Lesson 5 of the free MLOps course: who owns what across data science and engineering, where machine learning costs actually accumulate, the specific technical debt these systems create, and what governance requires.
📄️ 6. Recap and FAQ
Lesson 6 of the free MLOps course: a synthesis of the five lessons, a pre-launch checklist, twelve frequently asked questions, and a vocabulary you can use precisely.
📄️ 7. Quiz and attestation
Test your understanding of MLOps with a free 5-question quiz, corrected and explained: why a model is not software, reproducibility, drift monitoring, deployment patterns and rollback.