MLOps
The practices and tooling that take a machine-learning model from a notebook to reliable production: experiment tracking, model registries, feature stores, CI/CD for ML, serving, and drift monitoring. Builds on this library's docker-and-containers, kubernetes-fundamentals, harness, opentelemetry, and testing-in-software-engineering courses for the general infra/CI/testing/observability mechanics this course specializes for ML — see each for the foundation.
Get Yaaddi on Google Play Free. Install the app, then add this course from its Course Library.
View this course's repo on GitHub → — star it, open an issue, or suggest a fix.
Card type mix
- multiple choice: 145
- true false: 90
- select blank: 50
- multi select: 18
- match pairs: 18
- categorize: 6
- order: 5
- code fill: 3
- command output: 2
What's covered
- What MLOps Actually Solves
- The ML Lifecycle Loop
- Reproducibility Fundamentals
- Data Versioning
- Experiment Tracking Fundamentals
- MLflow Tracking
- Weights & Biases
- Model Registry & Lifecycle Stages
- Feature Store Fundamentals
- Feast in Practice
- Pipeline Orchestration Fundamentals
- Kubeflow Pipelines
- Airflow and Prefect for ML
- CI for ML: Data and Model Quality Gates
- Continuous Training and Retraining Triggers
- Packaging a Model for Deployment
- Model Serving Patterns
- Model Serving Tools
- Safe Rollouts: Canary and Shadow Deployment
- Monitoring ML Systems in Production
- Data Drift and Concept Drift
- Testing ML Systems
- Model Governance and Model Cards
- LLMOps: What Changes for Large Language Models
- Recap: Key Terms
Example cards
What gap does MLOps close?
Order the ML lifecycle loop steps.
Latest: Haiku review fixes: corrected factual errors and cards that did not match what their lesson teaches.