Exam Professional Machine Learning Engineer topic 1 question 279 discussion - ExamTopics

Problem

The question presents a scenario where an ML pipeline needs to be orchestrated across various Google Cloud services, triggered only by the addition of new files to a Cloud Storage bucket. The goal is to minimize costs.

Options

• A: Using Vertex AI Pipelines with a scheduler to periodically check for new files.
• B: Employing a Cloud Function with a Cloud Storage trigger to deploy a Cloud Composer DAG.
• C: Combining Vertex AI Pipelines with a Cloud Function triggered by Cloud Storage to deploy the pipeline.
• D: Utilizing a Cloud Composer DAG with a GCSObjectUpdateSensor to detect new files.

Suggested Solution

The suggested answer is C. This option combines the scalability of Vertex AI Pipelines with the event-driven nature of Cloud Functions, efficiently triggering pipeline execution only when necessary, thus minimizing cost.