Prepare and test your skills
Prepare and test your skills
Free sample question, answered and explained below. The Professional Cloud Architect practice bank has 2,335 more like it. Try 10 free: instant score, no account.
An enterprise machine learning team has defined an automated model retraining workflow using the Kubeflow Pipelines (KFP) SDK and compiled the pipeline specification into a YAML file stored in Cloud Storage.
The team wants to implement an event-driven architecture that triggers a new Vertex AI Pipeline execution whenever upstream data ingestion completes. The data ingestion service publishes a notification containing dynamic runtime parameter values.
Which architecture should you implement to trigger the Vertex AI Pipeline?
This option suggests using Cloud Build to continuously poll a Pub/Sub topic to compile and submit pipeline jobs.
The goal was to trigger automated pipeline compilation, unit testing, and deployment of pipeline YAML templates upon a Git commit or merge request in your source repository.
This configuration uses Cloud Scheduler to trigger HTTP endpoints on a fixed, time-based schedule (such as hourly or daily) using cron expressions.
The requirement was to schedule recurring continuous retraining runs on a predictable periodic schedule (e.g., every Sunday at midnight) rather than reacting immediately to completed data ingestion.
This option attempts to use Eventarc to monitor Cloud Storage and invoke Cloud Dataflow for compiling and executing the ML pipeline steps.
The workload involved transforming large volumes of raw streaming logs into clean tabular data prior to training.
aiplatform.PipelineJob.This pattern uses an event-driven Cloud Function subscribed to a Cloud Pub/Sub topic to automate pipeline execution. When upstream processes publish messages containing dynamic parameters, the function parses the message and uses the Vertex AI SDK (google-cloud-aiplatform) to launch a pipeline run.
base64 payload from the Pub/Sub event and extracts key-value parameter mappings.aiplatform.init() and executes aiplatform.PipelineJob.submit() pointing to the compiled YAML template in Cloud Storage.This approach aligns with Google Cloud best practices for reactive, serverless MLOps workflows by bridging asynchronous event sources to Vertex AI Pipelines without requiring continuously running virtual machines.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated Validated