# MLOps

> MLOps is the set of practices for deploying, monitoring and maintaining machine-learning models reliably in production — the ML equivalent of DevOps. It covers pipelines, versioning, testing, serving, monitoring and retraining.

*Source: https://www.lazlosoftwaresolution.com/glossary/mlops*

MLOps turns a model from a notebook experiment into a dependable production system: automated training/deploy pipelines, drift and quality monitoring, rollback, and reproducibility. It's what keeps AI reliable at scale.