With the October release of Cloudera Machine Learning, we have enabled the ability for CDP Machine Users to operate as Service Accounts in an ML Workspace. This feature is available both from within the ML Workspace UI and from CML APIs.
The Service Account plays a crucial role in moving ML/AI workloads into production for several reasons. They enable secure access to resources and services without tying these accesses to an individual's credentials. Service Accounts also can take on roles in an ML workspace enabling scenarios where a job, deployed model, or AI application can run under the identity of that Service Account. This enables MLOps and CI/CD workloads to run under the Service Account and not tied to a particular user. Furthermore, they offer traceability, as the actions performed using a service account can be audited and monitored, ensuring accountability and governance.