PMLE Practice Test 19 — 15 Questions

Question 1

A company uses Delta Lake on Dataproc for their data lake. They need to ensure ACID transactions and schema enforcement for data ingested from streaming sources. Which Delta Lake feature should they enable?

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

Schema enforcement. Delta Lake provides ACID transactions, schema enforcement, and time travel. Schema enforcement is key for streaming data to prevent data quality issues.

Answer

Delta Lake time travel

Answer

Delta Lake change data feed

Answer

Optimized write

Question 2

A company wants to use DVC for data versioning alongside their ML code in Git. Which TWO statements about DVC are correct? (Select 2)

Accepted Answer

DVC uses a separate .dvc file to track data versions.. DVC tracks data files by storing their hashes in Git and uses a remote storage for the actual data. It can integrate with cloud storage like GCS.

Answer

DVC stores the actual data files in Git.

Answer

DVC only works with AWS S3 as remote storage.

Answer

DVC replaces Git for code versioning.

Question 3

A team is training a model using historical data and wants to avoid data leakage when joining feature values from a feature store. The features include time-varying data like user activity counts. Which retrieval method should they use when creating a training dataset?

Accepted Answer

Use point-in-time correct retrieval with timestamp matching. Point-in-time correct retrieval joins features at the exact timestamp of each training row, ensuring no future data is used. This prevents data leakage. Other methods without timestamp handling introduce leakage.

Answer

Retrieve the latest feature values for each entity

Answer

Aggregate features over all historical data

Answer

Use random sampling of feature values

Question 4

A team uses Vertex AI Feature Store with an online store for low-latency serving. They need to support frequent updates to features (e.g., every minute) and require high write throughput (thousands of writes per second). Which online store type should they choose?

Accepted Answer

Bigtable online store. Bigtable online store is optimized for high write throughput and low-latency serving, suitable for frequently updated features. Optimized online store is better for read-heavy, static features.

Answer

Optimized online store

Answer

Firestore online store

Answer

Cloud SQL online store

Question 5

A data scientist is training a model using Vertex AI Experiments and wants to automatically log model parameters, metrics, and artifacts without modifying their training script. Which approach should they use?

Accepted Answer

Build a custom container with the Vertex AI SDK and MLflow installed, then set the VERTEX_AI_AUTOLOG environment variable to 1.. Vertex AI Experiments supports autologging via the Vertex AI SDK when using MLflow or Keras callbacks. Specifying a custom container with the SDK installed and enabling autologging allows automatic logging without code changes.

Answer

Use the gcloud beta ai experiments autolog command before running the training script.

Answer

Wrap the training code in a Kubeflow Pipelines component that calls the Vertex AI Experiments API.

Answer

Use the Vertex AI Python SDK to create a custom training job and manually log each parameter and metric in the training script.

Question 6

A data scientist wants to automatically generate model documentation that includes model purpose, training data, evaluation results, and intended use. Which tool should they use?

Accepted Answer

Model Cards in Vertex AI. Model Cards in Vertex AI provide a standardized framework for documenting models, automatically populated with metadata from Model Registry.

Answer

Vertex AI Workbench

Answer

Vertex AI Experiment

Answer

Cloud Datalab

Question 7

A logistics company wants to classify shipping documents into categories (invoice, packing slip, bill of lading) using a custom model with minimal code. They have labeled training images. Which Google Cloud service is most appropriate?

Accepted Answer

AutoML Vision for image classification. Document AI provides custom extractors but not classification. AutoML Vision can train a custom image classification model. Vertex AI AutoML Tables is for tabular data, not images. Cloud Vision API is for pre-built image analysis, not custom classification.

Answer

Vertex AI AutoML Tables

Answer

Document AI custom extractor

Answer

Cloud Vision API with label detection

Question 8

An ML team uses Vertex AI Workbench managed notebooks and wants to version their notebook code and collaborate using Git. Which THREE steps are required to set up Git integration? (Select 3)

Accepted Answer

Configure Git credentials (username/email) using git config.. To use Git in managed notebooks, you need to install git, configure credentials, and clone the repository.

Answer

Create a Cloud Source Repositories mirror of the GitHub repo.

Answer

Enable the Vertex AI Notebooks API for Git sync.

Question 9

A company wants to classify customer support emails into categories like 'billing', 'technical', or 'account'. They have labeled email text data. Which AutoML solution should they use?

Accepted Answer

AutoML Natural Language. AutoML Natural Language is designed for text classification. AutoML Vision is for images, AutoML Video for video, and AutoML Tables for tabular data.

Answer

AutoML Tables

Answer

AutoML Video

Answer

AutoML Vision

Question 10

A company uses Vertex AI Feature Store for feature engineering. They need to ensure point-in-time correctness to avoid data leakage during training. Which feature retrieval method should they use?

Accepted Answer

Use the offline store with point-in-time join using the `feature_view` with a timestamp column.. Point-in-time retrieval in Vertex AI Feature Store allows fetching feature values as they existed at a specific timestamp, preventing leakage.

Answer

Use the `get_features` API without specifying a timestamp.

Answer

Use BigQuery to manually join features with a sliding window.

Answer

Use the online store to retrieve the latest feature values.

Question 11

A company uses BigQuery as their data warehouse. They want to version datasets for ML experiments and be able to query snapshots at specific points in time. Which approach is most cost-effective and requires minimal operational overhead?

Accepted Answer

Create BigQuery table snapshots at key milestones.. BigQuery table snapshots are a built-in feature for point-in-time recovery and versioning at low cost (only storage changes).

Answer

Use BigQuery table clones to create copies of the data.

Answer

Export the table to Cloud Storage as Parquet and use DVC to version the files.

Answer

Use BigQuery time-travel (7-day window) to query historical data without snapshots.

Question 12

An organization uses Vertex AI Workbench user-managed notebooks and wants to enable collaboration where multiple data scientists can edit the same notebook simultaneously. Which configuration should they use?

Accepted Answer

Use a managed notebook instance (Colab Enterprise) and share the notebook via a link.. Vertex AI Workbench user-managed notebooks do not support real-time collaboration. Managed notebooks (now upgraded to Colab Enterprise) support simultaneous editing via integration with Colab.

Answer

Use Git to branch and merge notebooks, with a CI/CD pipeline to resolve conflicts.

Answer

Share the user-managed notebook instance by giving multiple users IAM roles to access the same JupyterLab instance.

Answer

Store the notebook in Cloud Storage and ask users to edit sequentially using gsutil cp.

Question 13

A team uses Vertex AI Workbench managed notebooks. They want to version control their notebook files and collaborate using Git. What is the best way to integrate Git?

Accepted Answer

Use the built-in Git integration in Vertex AI Workbench managed notebooks. Vertex AI Workbench managed notebooks support direct Git integration via the user interface, allowing clone, commit, and push operations.

Answer

Use Cloud Source Repositories only

Answer

Use gcloud source repos clone inside the terminal

Answer

Manually download notebooks and upload to GitHub via browser

Question 14

An organization needs to implement MLOps with standardized pipeline templates across multiple teams. Which Vertex AI feature should they use to create reusable pipeline components?

Accepted Answer

Vertex AI Pipelines. Vertex AI Pipelines allows creation of reusable pipeline components and templates that can be shared across teams, enabling standardization.

Answer

Vertex AI Experiments

Answer

Vertex AI Metadata

Answer

Vertex AI Workbench

Question 15

An ML team wants to implement data versioning for large datasets stored in Google Cloud Storage. They need to track changes over time and reproduce previous data states. Which tool is most appropriate?

Accepted Answer

DVC. DVC (Data Version Control) is designed specifically for versioning large datasets and ML models, working with cloud storage.

Answer

Cloud Storage Object Versioning

Answer

BigQuery table snapshots

Answer

Git LFS