Exam Professional Machine Learning Engineer topic 1 question 180 discussion - ExamTopics

Question:

The question describes a scenario where a machine learning pipeline is deployed using Vertex AI Pipelines and Cloud Build for CI/CD. The goal is to improve the process for deploying new pipeline implementations while minimizing the risk of production failures. Several options for structuring the CI/CD pipeline are provided.

Options:

• A. Build, test, and manually deploy to Artifact Registry and Vertex AI Pipelines.
• B. Build, deploy to pre-production, test in pre-production, then deploy to production.
• C. Build, test, deploy to pre-production, and upon successful pre-production run, deploy to production.
• D. Build, test, deploy to pre-production, rebuild, and then deploy to production.

Answer:

The correct answer is C. This option emphasizes a robust CI/CD process including automated testing in a pre-production environment before deploying to production. Options A, B, and D have flaws either in automation or in unnecessary steps (rebuilding in option D).