MLOps from training to Terraform, for free
Because training a model is the easy part; keeping it alive in production is the actual job.

What it does MLOps Zoomcamp is a free, nine-week curriculum from DataTalks.Club that teaches data professionals how to productionize machine learning services. The materials cover experiment tracking, pipeline orchestration, model deployment, monitoring, and infrastructure as code, with prerequisites that include Python, Docker, and at least one year of programming experience. All content is available for self-paced study, though no live cohort is currently scheduled for 2026.
The interesting bit The course gives monitoring the same weight as deployment, splitting it into web-service stacks (Prometheus, Evidently, Grafana) and batch-job stacks (Prefect, MongoDB, Evidently). It is also fully self-paced as of now, since the organizers have no 2026 cohort planned.
Key highlights
- Six structured modules plus a final end-to-end project, progressing from MLflow experiment tracking to Terraform infrastructure.
- Deployment coverage spans online strategies (Flask web services, AWS Kinesis/Lambda streaming) and offline batch scoring.
- Monitoring is split by architecture: Prometheus, Evidently, and Grafana for web services; Prefect, MongoDB, and Evidently for batch pipelines.
- Best-practices module includes testing, linting, CI/CD with GitHub Actions, and Infrastructure as Code.
- Community support lives in a dedicated Slack channel with posted guidelines and an FAQ document.
Caveats
- No live cohort is scheduled for 2026; self-paced learners get videos and Slack, but no instructor-led sessions.
- The prerequisites are steep for a “fundamentals” course: you need Python, Docker, command-line basics, prior ML exposure, and at least a year of programming.
- Module 3’s description is minimal—just “Workflow orchestration” and homework—so that section may be thinner than the others.
Verdict A solid syllabus to follow if you are a data scientist or engineer trying to bridge the gap between notebook prototypes and production systems. Look elsewhere if you need a plug-and-play deployment template; this is a semester-length curriculum, not a tool.
Frequently asked
- What is DataTalksClub/mlops-zoomcamp?
- Because training a model is the easy part; keeping it alive in production is the actual job.
- Is mlops-zoomcamp open source?
- Yes — DataTalksClub/mlops-zoomcamp is an open-source project tracked on heatdrop.
- What language is mlops-zoomcamp written in?
- DataTalksClub/mlops-zoomcamp is primarily written in Jupyter Notebook.
- How popular is mlops-zoomcamp?
- DataTalksClub/mlops-zoomcamp has 15k stars on GitHub and is currently accelerating.
- Where can I find mlops-zoomcamp?
- DataTalksClub/mlops-zoomcamp is on GitHub at https://github.com/DataTalksClub/mlops-zoomcamp.