PDE Practice Exam 54 — 20 Questions

Question 1

A team has multiple versions of a model and wants to manage them centrally, including tracking metadata and promoting versions to production. Which tool should they use?

Accepted Answer

Vertex AI Model Registry. Vertex AI Model Registry is the correct tool because it is purpose-built for centrally managing multiple model versions, tracking metadata (such as training parameters, evaluation metrics, and lineage), and promoting versions through stages like staging to production. Unlike generic storage or version control systems, it provides native integration with Vertex AI Pipelines and endpoints for controlled rollout and rollback.

Answer

Cloud Storage

Answer

BigQuery

Answer

GitHub

Question 2

A company stores data in a Cloud Storage bucket with versioning enabled. They want to automatically delete objects that are noncurrent (i.e., previous versions) after 30 days, and also delete the current version if it is older than 365 days. Which three Object Lifecycle Management conditions can be used together? (Choose three.)

Accepted Answer

age: 365. The `age` condition in Object Lifecycle Management specifies the number of days since object creation, and setting it to 365 will delete the current version when it is older than 365 days. This directly meets the requirement to delete current versions older than a year.

Answer

lastAccessTime: 30

Answer

daysSinceCustomTime: 30

Question 3

A data engineer needs to inspect a BigQuery table for sensitive data such as credit card numbers and email addresses before sharing it with a third party. The engineer also wants to de-identify the data by masking the sensitive columns. Which Google Cloud service should be used?

Accepted Answer

Cloud DLP. Cloud DLP (Data Loss Prevention) is the correct service because it is specifically designed to inspect, classify, and de-identify sensitive data such as credit card numbers and email addresses. It provides built-in infoType detectors for over 150 types of sensitive data and supports masking, tokenization, and other de-identification techniques. The engineer can use Cloud DLP to scan BigQuery tables and then apply transformations to mask the sensitive columns before sharing.

Answer

Dataplex

Answer

BigQuery column-level security

Answer

Data Catalog

Question 4

A data engineer needs to schedule a recurring batch load of CSV files from an on-premises SFTP server into BigQuery. The files are generated daily and need to be loaded into a partitioned table by date. Which THREE steps should the engineer take? (Choose THREE.)

Accepted Answer

Use Storage Transfer Service to copy files from the SFTP server to Cloud Storage every day. The Storage Transfer Service is designed to move data from on-premises sources (including SFTP servers) into Cloud Storage on a scheduled basis. This is the recommended first step for ingesting files from an external SFTP server into Google Cloud, as it handles the network transfer, retries, and scheduling natively without requiring custom code.

Answer

Create a Cloud Function triggered by Cloud Scheduler to load files directly from SFTP to BigQuery

Answer

Use BigQuery Data Transfer Service with SFTP as a source

Question 5

Which TWO actions can help reduce prediction latency for a Vertex AI endpoint?

Accepted Answer

Optimize the model architecture to reduce size. Optimizing the model architecture to reduce size directly decreases the computational load during inference, which lowers prediction latency. Smaller models require fewer floating-point operations (FLOPs) per prediction, enabling faster response times on Vertex AI endpoints.

Answer

Increase the number of features

Answer

Use a larger machine type with more vCPUs

Answer

Set min replicas to 0 to save cost

Question 6

A data engineer needs to monitor model performance over time for drift detection. What tool is specifically designed for this?

Accepted Answer

Vertex AI Model Monitoring. Vertex AI Model Monitoring is specifically designed to detect prediction drift and feature skew in deployed machine learning models. It continuously analyzes serving data against training data distributions and alerts when statistical metrics (e.g., Jensen-Shannon divergence, L-infinity distance) exceed configured thresholds, making it the correct tool for drift detection in the context of operationalizing ML models.

Answer

Cloud Monitoring

Answer

Cloud Logging

Answer

BigQuery ML

Question 7

Your team uses Cloud Composer to run Apache Airflow DAGs. One DAG uses a BigQueryInsertJobOperator to run a query and then uses BigQueryCheckOperator to verify the results. The DAG is failing intermittently because the query result is not ready when the check operator runs. How should you modify the DAG to ensure the check operator runs only after the query completes successfully?

Accepted Answer

Add a BigQueryTableExistenceSensor before the BigQueryCheckOperator to wait for the table to be created.. The BigQueryInsertJobOperator is synchronous by default, so the DAG should already wait for completion. However, if the check operator fails due to data not being written, adding a sensor operator (e.g., BigQueryTableExistenceSensor) between the two tasks ensures the table exists before checking.

Answer

Change the BigQueryInsertJobOperator to use deferrable mode to make it async.

Answer

Use a PythonOperator to call the BigQuery API directly and wait for the job to finish.

Answer

Increase the timeout of the BigQueryCheckOperator.

Question 8

A company uses Cloud Pub/Sub with pull subscriptions to process orders. The application requires at-least-once delivery and the ability to process orders in order per customer_id. Which THREE features should they configure? (Choose three.)

Accepted Answer

Configure a dead letter topic. A dead letter topic is correct because it allows undeliverable messages to be moved to a separate topic after all delivery attempts are exhausted, preventing message loss while still enabling at-least-once delivery semantics. This ensures that problematic messages do not block the processing of subsequent messages in the same ordering key group, which is critical when message ordering is enabled.

Answer

Use a push subscription with a HTTPS endpoint

Answer

Set the subscription's ackDeadline to 600 seconds

Question 9

A company wants to replicate a Cloud SQL (PostgreSQL) database to BigQuery in near real-time for analytics. The volume is about 10GB per day with frequent updates and deletes. They need to capture changes with low latency and ensure exactly-once delivery to BigQuery. Which approach should they use?

Accepted Answer

Use Cloud Data Fusion with a Debezium streaming source to capture CDC from Cloud SQL and a BigQuery sink with exactly-once mode.. Cloud Data Fusion with a Debezium streaming source provides native change data capture (CDC) from PostgreSQL, capturing inserts, updates, and deletes with low latency. The BigQuery sink in exactly-once mode ensures no duplicate records, meeting the requirement for near real-time analytics with frequent updates and deletes.

Answer

Export the entire database to Cloud Storage as CSV files every hour and load them into BigQuery using a load job with WRITE_TRUNCATE.

Answer

Use a Dataflow pipeline with JDBCIO to read from Cloud SQL every minute and write changes to BigQuery using upserts.

Answer

Use Cloud SQL's change data capture feature to write changes to a Pub/Sub topic and use a Dataflow pipeline to stream into BigQuery.

Question 10

Which THREE Google Cloud services are considered fully managed serverless data processing services? (Choose THREE.)

Accepted Answer

Cloud Functions. Cloud Functions is a fully managed serverless data processing service because it executes code in response to events without requiring any server provisioning or management. It automatically scales from zero to thousands of instances based on incoming requests, and you pay only for compute time used while your code runs. This makes it ideal for lightweight, event-driven data processing tasks such as transforming data in Cloud Storage or reacting to Pub/Sub messages.

Answer

Cloud Dataproc

Answer

Cloud Composer

Question 11

A data engineer is setting up CI/CD for a machine learning model using Cloud Build and Vertex AI. Which two components are essential? (Select 2)

Accepted Answer

Container Registry for model images. Container Registry (option B) is essential because it stores the Docker container images that encapsulate the trained model and its dependencies, which Cloud Build builds and pushes to the registry. Vertex AI Endpoints (option D) is essential because it provides the managed serving infrastructure to deploy the model image and expose it as a REST API for online predictions, enabling the CI/CD pipeline to automatically update the endpoint with new model versions.

Answer

Cloud Storage for datasets

Answer

Cloud Source Repositories

Answer

Cloud Functions for triggers

Question 12

A team wants to retrain a model weekly using new data stored in BigQuery. They want to minimize manual effort. Which approach should they use?

Accepted Answer

Create a Vertex AI Pipeline scheduled via Cloud Scheduler. Vertex AI Pipelines allow you to define a repeatable, automated ML workflow that can be triggered on a schedule via Cloud Scheduler. This minimizes manual effort by handling data extraction from BigQuery, model retraining, and deployment without human intervention, while also providing versioning and monitoring capabilities.

Answer

Use Cloud Scheduler to trigger a Cloud Function that retrains

Answer

Retrain manually in a notebook each week

Answer

Use Cloud Composer to orchestrate retraining

Question 13

Match each Google Cloud IAM role to its description.

Accepted Answer

BigQuery Admin: Full control over BigQuery resources, including datasets, tables, jobs, and permissions.. Predefined IAM roles in BigQuery control access at different levels. The BigQuery Admin role has full management, Data Editor can modify datasets/tables, Data Viewer can read data, and Job User can run queries. Common confusions mix these permissions.

Answer

BigQuery Admin: Can only view data and run queries.

Answer

BigQuery Data Viewer: Full control over BigQuery resources.

Question 14

A company deploys an ML model using Vertex AI Pipelines. They want to ensure reproducibility and traceability. Which TWO practices should they implement?

Accepted Answer

Pin all dependency versions. Pinning all dependency versions (Option A) ensures that every pipeline run uses the exact same library versions, eliminating variability from package updates. This is a fundamental practice for reproducibility because even a minor version bump can change model behavior or break code. In Vertex AI Pipelines, dependencies are typically specified in a `requirements.txt` or `Dockerfile`, and pinning them (e.g., `tensorflow==2.12.0`) guarantees consistent execution environments across runs.

Answer

Record dataset version using Vertex AI Dataset

Answer

Use custom containers for every step

Answer

Use Kubeflow Pipelines instead

Question 15

A data science team has trained a TensorFlow model for image classification and wants to deploy it to production with minimal latency. They have already exported the model as a SavedModel directory. Which service should they use to create an online prediction endpoint?

Accepted Answer

Vertex AI Endpoints. Vertex AI Endpoints is the correct service for deploying a TensorFlow SavedModel to an online prediction endpoint with minimal latency. It provides managed, autoscaling infrastructure optimized for real-time inference, including GPU/TPU support, request batching, and automatic health checking, which are essential for production deployment.

Answer

Cloud Functions

Answer

AI Platform Prediction (legacy)

Answer

Cloud Dataflow

Question 16

A company deploys a machine learning model on Vertex AI for online predictions. The model experiences intermittent spikes in traffic, causing latency increases. Which strategy should the company use to ensure consistent low latency during traffic spikes?

Accepted Answer

Enable autoscaling on the Vertex AI endpoint with appropriate min and max nodes.. Vertex AI endpoints support autoscaling, which dynamically adjusts the number of prediction nodes based on incoming traffic. By setting appropriate min and max nodes, the endpoint can scale up during traffic spikes to maintain low latency and scale down during low traffic to reduce costs. This ensures consistent performance without manual intervention.

Answer

Manually scale the deployed model to a larger machine type during peak hours.

Answer

Reduce the number of prediction nodes to minimize overhead.

Answer

Switch to batch prediction to handle all requests asynchronously.

Question 17

A company uses Vertex AI Feature Store for serving features to both training and prediction. The team notices that predictions made shortly after training use different feature values, causing a training-serving skew. What is the most effective way to prevent this skew?

Accepted Answer

Configure the Feature Store to use point-in-time lookup using the training timestamp. Point-in-time lookup ensures that feature values used during training are exactly the same as those used during prediction by retrieving the feature value as it existed at the training timestamp. This directly addresses training-serving skew caused by time-dependent feature changes, which is a common issue in Vertex AI Feature Store when features are updated after training.

Answer

Retrain the model more frequently to adapt to the new feature distributions

Answer

Use batch prediction instead of online prediction to ensure consistent features

Answer

Ensure that the training and prediction environments use identical compute resources

Question 18

An MLOps team manages a pipeline that retrains an XGBoost classifier weekly using BigQuery data. The pipeline is orchestrated with Cloud Composer and deploys the new model to Vertex AI Endpoint if validation metrics (AUC > 0.9) are met. Over the past month, the deployed model's AUC has dropped from 0.95 to 0.88, despite the training pipeline consistently reporting AUC > 0.9. Which THREE steps should the team take to diagnose and fix this issue?

Accepted Answer

Add a canary deployment step where new model version receives a small percentage of traffic before full rollout.. A canary deployment allows the team to gradually roll out the new model to a small percentage of traffic, enabling early detection of performance degradation in production before a full rollout. This step directly addresses the discrepancy between training metrics and live performance by exposing the model to real-world data patterns that may differ from the training set. In Cloud Composer and Vertex AI, canary deployments can be implemented by routing a fraction of requests to the new model version and monitoring its AUC in real time.

Answer

Review the training pipeline's hyperparameter tuning configuration to ensure it is not overfitting to stale data.

Answer

Retrain the model using a longer training history to include older data that may still be relevant.

Question 19

Match each BigQuery feature to its description.

Accepted Answer

Slots: Virtual CPUs used for parallel query processing.. BigQuery uses slots for compute, partitioning and clustering for storage optimization, and materialized views for query performance. Common confusions include mixing clustering with materialized views or partitioning with BI Engine.

Answer

Partitioning: Allows you to share BI dashboards directly from BigQuery without separate tools.

Answer

Clustering: Precomputes query results to reduce execution time for repeated queries.

Question 20

Your company deploys a classification model on Vertex AI for online predictions. The model is an XGBoost model trained on tabular data with 500 features. The endpoint uses a single n1-standard-4 node. After deployment, users report that predictions take 8-10 seconds on average, while the required SLA is under 2 seconds. You have already verified that the model is not large (under 100 MB) and the input data size is small. The endpoint does not scale automatically. Which action should you take to reduce latency to meet the SLA? A) Change the machine type to n1-highcpu-4 to prioritize compute over memory. B) Enable autoscaling by setting min replicas to 2 and max replicas to 5. C) Switch to a custom container that preloads the model into memory. D) Reduce the number of features by half.

Accepted Answer

Enable autoscaling by setting min replicas to 2 and max replicas to 5.. (Enable autoscaling) is correct because the current single-node endpoint is experiencing high latency due to request queuing or concurrency limits. Autoscaling with a minimum of 2 replicas distributes the load, reducing per-request latency. Option A (changing machine type) does not address concurrency; the n1-standard-4 already provides sufficient resources for the small model. Option B (reducing features) may degrade accuracy without guaranteed latency improvement. Option C (custom container) is unnecessary because Vertex AI already preloads models into memory by default.

Answer

Change the machine type to n1-highcpu-4 to prioritize compute over memory.

Answer

Reduce the number of features by half.

Answer

Switch to a custom container that preloads the model into memory.