Google Professional Machine Learning Engineer Question 312
Select 3Google Cloud PlatformYour organization has implemented several machine learning models that are used across different teams. You are tasked with organizing a model registry on Google Cloud to streamline model management. Which practices should you follow to ensure effective model versioning and deployment tracking?
- A
Use Vertex AI Model Registry to store and version models centrally.
- B
Assign unique version tags to models based on changes in data, architecture, or hyperparameters.
- C
Manually track model versions in a shared spreadsheet to ensure transparency across teams.
- D
Implement metadata tracking for each model version, including training data source, evaluation metrics, and deployment status.
- E
Avoid storing older model versions to reduce storage costs in the registry.
Show answer and explanation
Correct answers: A, B, D
Explanation
Organizing a model registry involves leveraging tools like Vertex AI Model Registry for centralized model management, versioning models effectively, and tracking metadata for each version. These practices ensure scalability, reproducibility, and compliance, which are vital for managing machine learning workflows in professional environments. Manually tracking models or avoiding the storage of older versions can lead to inefficiencies and potential risks.
- A. Correct.
Using Vertex AI Model Registry is a best practice for centralizing model management in a scalable and organized way. It supports versioning, metadata tracking, and integration with other Google Cloud services.
- B. Correct.
Assigning unique version tags ensures that changes to data, architecture, or hyperparameters are traceable, which is essential for reproducibility and debugging.
- C. Incorrect.
Manually tracking model versions in spreadsheets is error-prone and not scalable. Automated solutions like Vertex AI Model Registry are recommended for professional ML workflows.
- D. Correct.
Metadata tracking, such as training data source, evaluation metrics, and deployment status, is critical for understanding model performance and ensuring compliance with organizational policies.
- E. Incorrect.
Avoiding the storage of older model versions is not recommended. Older versions are often necessary for debugging, rollback strategies, and compliance requirements.