Training Trains Internship
Student login
← MLOps learning path
MLOps · Module 6 of 10

CI/CD for machine learning

Test and deliver ML code and models automatically.

📖 Key ideas

  1. CI runs tests on code changes: unit tests, data checks, linting.
  2. CD deploys approved models automatically.
  3. Test data schemas, feature logic and model performance thresholds.
  4. Use GitHub Actions, GitLab CI or Jenkins.
  5. Promote models through staging before production.
💡 Real-life example: CI/CD for ML is a quality gate at an airport: only checked bags (models) board the plane (production).
CIContinuous Integration
CDContinuous Delivery/Deployment
Quality gateThreshold a model must pass
StagingPre-production environment