Domain 5: Automating and Orchestrating ML Pipelines¶
Official exam weight: approximately 18%
Official objectives¶
5.1 Develop end-to-end ML pipelines¶
- Validate data and models.
- Build and orchestrate pipelines using managed or unmanaged services, templates, or custom solutions.
- Understand Agent Platform Pipelines, Managed Service for Apache Airflow, and Ray on Agent Platform.
- Ensure preprocessing is consistent between training and serving.
5.2 Automate model retraining¶
- Determine an appropriate retraining policy.
- Deploy models through continuous integration, continuous delivery, and continuous training (CI/CD/CT) pipelines, including Cloud Build.
Guided labs to build¶
| Status | Lab | Objective |
|---|---|---|
| Planned | Build a reusable end-to-end ML pipeline | 5.1 |
| Planned | Add data and model validation gates | 5.1 |
| Planned | Orchestrate preprocessing, training, and evaluation | 5.1 |
| Planned | Keep training and serving preprocessing consistent | 5.1 |
| Planned | Schedule or trigger model retraining | 5.2 |
| Planned | Implement a CI/CD/CT workflow with Cloud Build | 5.2 |
Google Skills¶
Open the official PMLE learning path for the current pipeline and MLOps activities.
Exam focus¶
Understand the distinction between a standalone training job and a repeatable multi-step pipeline. Be able to choose managed or self-managed orchestration, determine when retraining should occur, and explain how CI/CD/CT reduces manual and inconsistent model releases.
📬 DevopsPilot Weekly — Learn DevOps, Cloud & Gen AI the simple way.
👉 Subscribe here