Prepare and test your skills

Prepare and test your skills

Free sample question, answered and explained below. The Professional Data Engineer practice bank has 1,353 more like it. Try 10 free: instant score, no account.
A data engineering team is deploying automated machine learning pipelines with BigQuery ML and Vertex AI. To satisfy strict governance and reproducibility requirements across iterative model runs, the team must implement an architecture that:
Which combination of Google Cloud services and practices should the team implement?
This architecture combines BigQuery table snapshots, Vertex ML Metadata (integrated through Vertex AI Experiments), and native BigQuery ML monitoring functions like ML.VALIDATE_DATA_SKEW to build a complete, reproducible, and verifiable ML lifecycle.
ML.VALIDATE_DATA_SKEW function compares incoming production serving data against these precomputed training statistics to detect distribution shifts without requiring raw training data reloading.Combining native table snapshots with managed ML metadata and precomputed statistical skew evaluation provides a fully integrated, enterprise-grade MLOps workflow without custom infrastructure overhead.
This approach relies on standard BigQuery views for data retrieval, uses Artifact Registry as a repository for pipeline parameters and code, and implements ML.VALIDATE_DATA_DRIFT for production monitoring.
ML.VALIDATE_DATA_DRIFT evaluates differences between two distinct windows of production serving data over time, rather than validating production serving data against the baseline training distribution.The goal were simply to compare consecutive production serving windows (e.g., comparing day-over-day production inference batches) while versioning container build environments in CI/CD pipelines.
This proposal creates materialized BigLake tables for each run, manages hyperparameters inside Vertex AI Feature Store, and relies on Vertex Explainable AI feature attributions to prevent distribution skew.
The primary requirement were low-latency online feature retrieval with explainability dashboards for deployed model predictions.
This option exports tabular datasets to Cloud Storage as CSV files, uses Cloud Logging custom metrics for workflow logging, and calls ML.DESCRIBE_DATA to inspect production inputs.
ML.DESCRIBE_DATA computes univariate descriptive statistics on a single dataset; it does not perform automated statistical distance comparison or skew testing between training and serving datasets.You needed a lightweight method to compute summary statistics for exploratory data analysis without performing comparative skew checks against stored model baselines.
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