Glossary · M
MLOps
Machine Learning Operations
Practices for operating machine learning systems
What it means
MLOps covers the development, deployment, monitoring and maintenance of machine learning systems. It extends software delivery practices to include model and dataset versioning, evaluation, reproducible training, drift monitoring and controlled model updates.
What it means for a business
A delivery-time prediction model should be monitored as products, routes and customer behaviour change. Versioned models and evaluation data help a team decide whether to retrain, adjust or roll back a model when performance no longer meets the agreed criteria.