Google Professional Machine Learning Engineer Question 167
Select 3Google Cloud PlatformYou are a Machine Learning Engineer working on a Google Cloud-based ML pipeline. Your team develops and maintains model training and serving code in a centralized Git repository. You are tasked with integrating this repository with your Google Cloud environment to automate CI/CD processes for training and deploying models. Which steps should you take to achieve this integration effectively?
- A
Use Cloud Build triggers to automatically build and deploy code changes when new commits are pushed to the repository.
- B
Use Vertex AI Model Registry to directly pull code from the Git repository for model training.
- C
Configure service account permissions to allow Cloud Build to access the Git repository securely.
- D
Set up a CI/CD pipeline using Artifact Registry to store containerized images of your training and serving code.
- E
Manually download the code from the Git repository and upload it to Cloud Storage for each new release.
Show answer and explanation
Correct answers: A, C, D
Explanation
To effectively integrate a Git repository with Google Cloud for CI/CD, you need to automate build and deployment processes using tools like Cloud Build triggers. Additionally, secure access to the repository is essential, which requires configuring service account permissions. Finally, storing containerized images in Artifact Registry ensures that the CI/CD pipeline can deploy code efficiently and reliably. These steps align with best practices for managing ML pipelines in a cloud environment.
- A. Correct.
Using Cloud Build triggers ensures that code changes in the repository automatically initiate builds and deployments, making it a key part of integrating source repositories with Google Cloud.
- B. Incorrect.
Vertex AI Model Registry is used to manage and serve trained ML models but cannot directly pull code from a Git repository for training purposes.
- C. Correct.
Service accounts with appropriate permissions are required for Cloud Build to securely access the Git repository and perform CI/CD tasks.
- D. Correct.
Artifact Registry is used to store containerized images, which are a common format for deploying code in CI/CD pipelines on Google Cloud.
- E. Incorrect.
Manually downloading and uploading code is not an automated or scalable solution and does not align with best practices for CI/CD pipelines.