MLOps

Definitions

1. A set of practices combining machine learning, DevOps, and data engineering to reliably deploy and maintain ML systems in production. MLOps covers the full lifecycle: experiment tracking, model versioning, automated retraining, CI/CD for models, monitoring for data drift and model degradation, and reproducibility of results.

   In plain English: The practice of managing AI models the same way software teams manage code — with version control, automated testing, and monitoring — so models stay reliable in production.

   Example: Their MLOps pipeline automatically retrains the fraud model weekly on fresh data, runs validation tests, and promotes it to production only if precision stays above 95%.

Related Terms

• CI/CD
• DevOps
• feature engineering
• feature store
• model serving
• CUDA
• Puppet