Project

For the cross-course logistics, see Final Project (Zoomcamp Logistics). That page covers attempts, deadlines, peer review, and certification mechanics.

This page covers what is specific to the MLOps Zoomcamp.

Goal

The project is an end-to-end MLOps pipeline. You pick a problem and a dataset, train a model, and then apply the practices from the course to put it into production and operate it.

What the project should demonstrate

A complete project applies the modules together:

• A clear problem description.
• Experiment tracking and a model registry (Module 2).
• A training pipeline that can be run reproducibly (Module 3).
• Model deployment, as a web service, streaming, or batch (Module 4).
• Model monitoring (Module 5).
• Best practices: tests, linting, CI/CD, and Infrastructure as Code where appropriate (Module 6).
• Reproducibility: clear instructions and pinned dependencies so a reviewer can run it.

Where to find the details

The 07-project folder in the repository contains the project instructions and the evaluation criteria. The rubric there is the authoritative description of what is graded.

Certificate

Completing the project during a live cohort is how you earn the certificate. Since there is no live 2026 cohort, the project is currently for self-assessment and portfolio building. See Certification for how the certificate is issued, the cross-course rules, and how to add it to LinkedIn.