Understanding the need for continuous integration and continuous deployment

Continuous integration (CI) and continuous deployment (CD) enable continuous delivery to the ML service. The goal is to maintain and version the source code used for model training, enable triggers to perform necessary jobs in parallel, build artifacts, and release them for deployment to the ML service. Several cloud vendors enable DevOps services that can be used for monitoring ML services, ML models in production, and orchestration with other services in the cloud.

Using CI and CD, we can enable continuous learning, which is critical for the success of an ML system. Without continuous learning, an ML system is destined to end up as a failed PoC (Proof of Concept). We will delve into the concepts of CI/CD and implement hands-on CI and CD pipelines to see MLOps in play in the next chapter.