Course Wiki

MLOps Zoomcamp

DataTalks.Club's free MLOps course: experiment tracking, orchestration, deployment, monitoring, and engineering best practices around one NY Taxi use case.

MLOps Zoomcamp is DataTalks.Club’s free course on productionizing machine learning services. It teaches the fundamentals of MLOps — experiment tracking, orchestration, deployment, and monitoring — through structured modules, hands-on workshops, and a final project that wires everything into one end-to-end pipeline. The running example is the NY Taxi dataset.

All materials are open source in the course repository, with videos on YouTube and a course FAQ. The course is currently self-paced only — no live cohort is scheduled for 2026, so certificates are paused until one runs again.

Curriculum

All modules, each with its lesson-level notes:

• Module 1: Introduction
• Module 2: Experiment Tracking and Model Management
• Module 3: Orchestration and ML Pipelines
• Module 4: Model Deployment
• Module 5: Model Monitoring

• Module 6: Best Practices

• Module 1: Introduction — what MLOps is and the MLOps maturity model.
• Module 2: Experiment Tracking and Model Management — experiment tracking with MLflow, model saving and loading, and the model registry.
• Module 3: Orchestration and ML Pipelines — turning the training notebook into a script and wrapping it in an orchestrator (workflow orchestration, e.g. Prefect or Airflow).
• Module 4: Model Deployment — the three deployment shapes: web services with Flask and Docker, streaming with AWS Kinesis and Lambda, and batch scoring (model deployment), pulling models from the model registry.
• Module 5: Model Monitoring — reference datasets and Evidently metrics and dashboards, web service monitoring with Prometheus and Grafana, and batch job monitoring with Prefect, MongoDB, and Evidently (model monitoring).
• Module 6: Best Practices — testing Python code with pytest, integration tests with docker-compose and LocalStack, linting and formatting, Git pre-commit hooks, CI/CD with GitHub Actions, and infrastructure as code with Terraform.
• Final project — an end-to-end MLOps pipeline integrating every module: tracked experiments, a training pipeline, a deployed model, monitoring, and best practices.

Looking for a specific concept? The Concept Glossary links every concept this course teaches to the exact modules and lessons that cover it.

Who it is for

The course targets data scientists, ML engineers, and software engineers who need to put models into production and operate them reliably. Prerequisites are real: Python, Docker basics, command-line comfort, prior machine learning experience (for example through ML Zoomcamp), and about a year of programming.