An automated pipeline that tests every code change as it is merged and releases the changes that pass, so software ships in small, frequent steps.
For AI systems, CI/CD pipelines extend into MLOps pipelines. Alongside code deployments, they automate data drift monitoring, model retraining, and evaluation against defined metrics. A consistent CI/CD process keeps delivery dependable as the architecture grows more complex.
See also MLOps, Data Drift, Technical Drag, Air-Gapped Environment, Role-Based Access Control and Zero Trust Architecture.