PMLE Practice Exam 55 — 20 Questions

Question 1

You are an ML engineer at a global e-commerce company. Your team has developed a deep learning model for product recommendation that runs on Vertex AI Prediction. The model is deployed on a single n1-highmem-2 instance (CPU only) with autoscaling enabled (min replicas=1, max replicas=10). During Black Friday, traffic spikes to 1000 requests per second (QPS), and you observe that latency increases from 50ms to over 5000ms, and many requests time out. You check the monitoring dashboard and see that CPU utilization is at 100% on the single instance, and autoscaling is not triggering quickly enough. The team has a budget for this service and wants to handle the spike without compromising latency. What should you do?

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

Accepted Answer

Switch to GPU instances (e.g., n1-standard-4 with T4) and set min replicas=2 with autoscaling up to 10. Switching to GPU instances (n1-standard-4 with T4) offloads compute-intensive recommendation model inference to GPUs, significantly reducing per-request latency. Setting min replicas=2 ensures that at least two instances are always warm, reducing cold-start delays and allowing autoscaling to handle traffic spikes more responsively. This combination addresses both the CPU bottleneck and the slow scaling trigger, keeping latency under 50ms even at 1000 QPS.

Answer

Increase min replicas to 5 to keep warm instances

Answer

Set min replicas=1 and max replicas=5 to control cost

Answer

Increase max replicas to 20 and keep CPU instances

Question 2

A data engineer is troubleshooting a Vertex AI Endpoint that serves a large BERT model. After deployment, many prediction requests fail with 'Out of Memory' errors. The machine type is n1-standard-8 (30 GB memory) with no accelerator. Which action will most likely resolve the issue?

Accepted Answer

Change the machine type to n1-highmem-16 (104 GB memory).. The n1-standard-8 machine type provides 30 GB memory, which is likely insufficient for a large BERT model due to intermediate tensors during inference exceeding this limit. Upgrading to n1-highmem-16 (104 GB memory) increases available memory, directly addressing OOM errors. Option B is incorrect because batch prediction is for offline processing, not real-time, and may still encounter memory issues. Option C is incorrect because adding a GPU does not increase system memory; it offloads computation but OOM due to memory still occurs. Option D is incorrect because quantization reduces model size but may not eliminate memory spikes from intermediate tensors, and it could affect accuracy.

Answer

Use batch prediction instead of online prediction.

Answer

Add a GPU accelerator (e.g., NVIDIA T4) to offload computation.

Answer

Quantize the model from FP32 to INT8.

Question 3

After deploying a new version of a model to a Vertex AI Endpoint, the team notices that predictions are still returning results from the old version. The deployment command used a traffic split of 100% to the new version. What is the most likely cause?

Accepted Answer

The traffic split was not properly updated; the endpoint is still routing 100% to the old version.. The most likely cause is that the traffic split was not properly updated. Although the deployment command specified 100% traffic to the new version, the actual traffic split may not have been applied correctly, so the endpoint continues routing all traffic to the old version. Option A is incorrect because an identical artifact would not cause predictions to come from the old version. Option C is incorrect because a health check failure would prevent the new version from serving, but the deployment would have failed, not silently roll back. Option D is incorrect because Vertex AI does not cache predictions on the client side.

Answer

The model artifact uploaded was identical to the old version.

Answer

The new model version failed health checks and was automatically rolled back.

Answer

The prediction client is caching the old model response.

Question 4

Which THREE actions should be taken to manage model versions effectively?

Accepted Answer

Use Vertex AI Model Registry. Vertex AI Model Registry is a centralized repository that tracks, versions, and manages ML models. It enables you to organize models, assign aliases (like 'champion' or 'experiment'), and control deployment, ensuring reproducibility and governance across the model lifecycle.

Answer

Delete old versions immediately

Answer

Use the same model name for all versions

Question 5

A team is architecting a low-code ML system for real-time predictions with AutoML. Which THREE considerations are critical for production?

Accepted Answer

Enable autoscaling for the endpoint. A is correct because autoscaling ensures the prediction endpoint can handle variable request loads without manual intervention, which is critical for production real-time systems. In Vertex AI, you can configure autoscaling with a target utilization level (e.g., 60%) to automatically adjust the number of compute nodes based on incoming traffic, preventing both over-provisioning and latency spikes.

Answer

Use a custom container for prediction

Answer

Use global model endpoints for low latency everywhere

Question 6

A company uses Vertex AI Pipelines to orchestrate an AutoML tabular training step followed by a BigQuery ML evaluation step. The pipeline fails because the output of the AutoML step (a model resource name) is not being passed to the BigQuery step. What is the most likely cause?

Accepted Answer

The AutoML training component is implemented as a Python function without proper artifact input/output annotations. In Vertex AI Pipelines, when using the Kubeflow Pipelines SDK, components must explicitly declare their inputs and outputs using type annotations (e.g., `Input[Model]`, `Output[Model]`) or via `@component` decorators with `outputs` specified. If the AutoML training step is implemented as a plain Python function without these annotations, the pipeline framework cannot serialize and pass the model resource name as an artifact to the downstream BigQuery ML evaluation step. This causes the pipeline to fail because the BigQuery step receives no valid model reference.

Answer

The pipeline is using a custom pipeline root but the model is in a different region

Answer

The Vertex AI Pipeline Runner does not have permission to access AutoML models

Answer

The BigQuery ML evaluation component requires a service agent with Cloud SQL access

Question 7

A company serves a PyTorch model using a custom container on Vertex AI Prediction. They notice that after a few hours, the endpoint returns 502 errors. The logs show 'Out of memory' errors. The container has a memory limit of 4GB, and the model loads a 3GB vocabulary file. What is the most likely cause and best fix?

Accepted Answer

Load the vocabulary file once at startup and reuse it.. The 502 errors and 'Out of memory' errors indicate that the container is running out of memory during inference. Since the model loads a 3GB vocabulary file, and the container has only 4GB of memory, loading this file repeatedly for each prediction request (e.g., inside the prediction handler) would quickly exhaust memory. The correct fix is to load the vocabulary file once at container startup and reuse it across all requests, which is a standard best practice for serving models with large static assets.

Answer

Increase the container memory to 8GB.

Answer

Increase the number of replicas to distribute load.

Answer

Switch to Vertex AI Batch Prediction.

Question 8

A team is using Vertex AI Pipelines to deploy a model. They have a component that evaluates the model and produces a ClassificationMetrics artifact. The pipeline should deploy the model only if the precision is greater than 0.9. They use dsl.If to check the metric. However, the condition always evaluates to False. What is the most likely cause?

Accepted Answer

The ClassificationMetrics artifact is not accessible in the condition context.. In Vertex AI Pipelines, `ClassificationMetrics` artifacts are not directly accessible as primitive values within the `dsl.If` condition context. The `dsl.If` condition can only evaluate pipeline parameters or primitive outputs (like strings, integers, floats) that are explicitly passed as pipeline-level parameters or task outputs. A `ClassificationMetrics` artifact is a complex object that must be parsed or have its specific metric values extracted (e.g., via a custom component or `dsl.Metrics`) before they can be used in a conditional check.

Answer

The precision value is stored as a float but the condition expects a string.

Answer

The evaluation component did not output the metric correctly.

Answer

The dsl.If block is placed incorrectly in the pipeline definition.

Question 9

Drag and drop the steps to implement a CI/CD pipeline for ML models using Cloud Build and Vertex AI in the correct order.

Accepted Answer

Configure the trigger, then Define the pipeline, then Commit to trigger training and deployment. In a CI/CD pipeline for ML models using Cloud Build and Vertex AI, the correct order is: first, configure the trigger (e.g., a Cloud Build trigger linked to a source repository), then define the pipeline (e.g., using Cloud Build configuration file or Vertex AI Pipelines), and finally commit code changes to trigger the automated training and deployment. This ensures that the trigger is set up to detect changes and the pipeline is defined before execution.

Answer

Define the pipeline, then Configure the trigger, then Commit to trigger training and deployment

Answer

Commit, then Configure the trigger, then Define the pipeline

Answer

Configure the trigger, then Commit, then Define the pipeline

Question 10

A team deploys a real-time model using a custom container on Vertex AI Prediction. The container is large (5 GB) and cold starts are causing latency spikes. The endpoint is configured with `min_replica_count=0` to reduce cost. The team wants to keep the cost low while reducing cold starts. What is the best approach?

Accepted Answer

Set `min_replica_count=1` to keep at least one replica always warm.. Setting min_replica_count=1 ensures that at least one replica is always running, eliminating cold starts for most requests. The container is large, but keeping a replica warm avoids the cold start latency. Option B is incorrect because using a prebuilt container may reduce image size but does not directly address cold starts; it might help with download time but not eliminate cold starts. Option C is incorrect because container memory optimization does not significantly reduce startup time for a large container. Option D is incorrect because provisioning a Persistent Disk (SSD) speeds up image download but does not fully eliminate cold start latency; the replica still needs to initialize.

Answer

Use a prebuilt container for the model framework to reduce image size.

Answer

Enable container memory optimization to reduce startup time.

Answer

Provision a Persistent Disk (SSD) for the container image to speed up download.

Question 11

Which TWO actions can help reduce prediction latency for a model deployed on Vertex AI Endpoint without changing the model architecture?

Accepted Answer

Attach a GPU accelerator to the endpoint's machine type.. Options B and C are correct. Option B (GPU accelerator) can significantly speed up inference for deep learning models. Option C (model quantization) reduces model size and inference time. Option A (increasing batch size) increases latency per request. Option D (multiregion deployment) reduces network latency but not prediction latency. Option E (smaller machine type) may increase latency.

Answer

Increase the batch size of prediction requests.

Answer

Deploy the model in multiple regions and use global load balancing.

Answer

Use a smaller machine type to reduce complexity.

Question 12

A machine learning engineer notices that a model served on Vertex AI Endpoints returns predictions that are consistently 20% slower during the first request after idle (cold start). They are using automatic scaling with min replicas=1. What is the most likely cause and best solution?

Accepted Answer

Vertex AI endpoint warm-up time; set min replicas to 2 to always keep a warm instance. With min replicas=1, a single instance is always running, but it may become idle and require a warm-up period to reload the model into memory, causing the first request to be slower. Setting min replicas to 2 ensures at least one instance remains fully warm to handle requests immediately, reducing cold start latency. Option B is incorrect because health checks measure readiness but do not actively warm the instance; they only prevent traffic until the instance is ready, which doesn't address the delay. Option C is incorrect because network latency would affect all requests consistently, not just the first after idle. Option D is incorrect: a smaller container image reduces initial load time but does not prevent the warm-up delay after idle, and the main issue is the model or endpoint warm-up, not container size.

Answer

Model loading time is high; enable health checks to warm the instance

Answer

Network latency; deploy to a different region

Answer

Container initialization delay; use a smaller container image

Question 13

You deploy a PyTorch model to Vertex AI Online Prediction. After deployment, you observe that inference latency is approximately 300ms per request, but the desired SLA is under 100ms. The model uses a custom container with CPU only. Which action is most likely to reduce latency to the target?

Accepted Answer

Deploy the model on a machine with a GPU accelerator.. Enabling GPU acceleration can significantly speed up inference for deep learning models. Adding more CPU instances may help with throughput but not per-request latency. Switching to batch prediction changes the use case, and using a smaller instance type might reduce latency if the model is small, but GPU is more impactful.

Answer

Switch from online prediction to batch prediction.

Answer

Increase the min_replica_count to ensure more instances are always available.

Answer

Use a smaller machine type with less CPU to reduce overhead.

Question 14

A team uses Vertex AI Prediction with a custom container. They want to perform canary deployments by sending 5% of traffic to a new model version. Which method should they use?

Accepted Answer

Use the Vertex AI Model Registry and configure traffic splitting on the endpoint. Vertex AI endpoints support traffic splitting between multiple model versions deployed to the same endpoint. This allows you to send 5% of traffic to a new version while the remaining 95% goes to the current version, enabling canary deployments without managing separate load balancers. Option B is incorrect because deploying two separate endpoints with a load balancer is not the standard method in Vertex AI; traffic splitting is natively supported on a single endpoint, making it simpler and more manageable.

Answer

Create a new endpoint with manual traffic splitting

Answer

Deploy two separate endpoints and use a load balancer

Answer

Use Cloud Run for serving with gradual rollout

Question 15

Which TWO statements are true about canary deployments for Vertex AI endpoints?

Accepted Answer

You can roll back a canary by resetting traffic to 0% for the new version.. Traffic splitting in Vertex AI endpoints allows you to gradually shift traffic from 1% to 100% to a new version. You can adjust the percentage at any time, including increasing it up to 100% for full rollout. To roll back, set the new version's traffic to 0%, sending all traffic to the old version. Canary deployments help test a new version with a small traffic percentage before full rollout, and automated rollback can be triggered based on monitoring metrics.

Answer

Canary deployments are only supported for custom containers, not prebuilt frameworks.

Answer

Canary deployments require the use of Vertex AI Model Registry.

Answer

Once a canary receives 50% traffic, you cannot increase it further.

Question 16

Your team manages multiple ML models on Vertex AI. You need to implement a centralized monitoring solution to track model performance over time. Which TWO approaches should you consider? (Choose two.)

Accepted Answer

Create Cloud Monitoring dashboards and alerts based on Vertex AI metrics.. Cloud Monitoring provides centralized dashboards and alerting for Vertex AI metrics such as prediction latency, request count, and error rates, enabling you to track model performance over time without additional infrastructure. Option D is correct because Vertex AI Model Monitoring is purpose-built to detect training-serving skew and feature drift by comparing serving data distributions to training data, which is essential for maintaining model performance in production.

Answer

Store all prediction logs in BigQuery and analyze using SQL.

Answer

Use Cloud Source Repositories to track model code versions.

Answer

Enable Cloud Billing budgets to track cost per model.

Question 17

A company has a large dataset of 1 million unlabeled images for object detection. They want to use AutoML Vision but need to minimize labeling effort. Which strategy should they use?

Accepted Answer

Use Vertex AI Active Learning to choose a subset for labeling. Vertex AI Active Learning is the correct strategy because it intelligently selects the most informative unlabeled images for human labeling, maximizing model accuracy while minimizing labeling effort. This approach uses the model's uncertainty to prioritize data points that will most improve performance, making it ideal for large datasets where manual labeling of all images is impractical.

Answer

Apply data augmentation techniques to increase dataset size

Answer

Manually label all 1 million images

Answer

Train a custom object detection model on unlabeled data with unsupervised learning

Question 18

You need to reduce the cost of training a large model on Vertex AI while maintaining fault tolerance. Which THREE actions should you take? (Choose 3)

Accepted Answer

Use spot VMs. Spot VMs are cheaper but can be preempted; using checkpointing and saving to a durable storage (like GCS) allows recovery. Also, using a single node with multiple GPUs may be cheaper than multiple nodes.

Answer

Use MultiWorkerMirroredStrategy

Answer

Use TPUs instead of GPUs

Question 19

Which TWO of the following are benefits of using BigQuery ML for low-code model development?

Accepted Answer

Train models directly on data in BigQuery without moving it. BigQuery ML allows you to train machine learning models using SQL directly on data stored in BigQuery, eliminating the need to export or move data to a separate environment. This reduces data transfer latency, simplifies security governance, and leverages BigQuery's native storage and compute separation.

Answer

Automatic feature engineering and hyperparameter tuning

Answer

Built-in model explainability for all model types

Answer

Support for image classification tasks

Question 20

A team has a Vertex AI pipeline that includes a container component for data preprocessing. The team notices that the component is re-executed every time the pipeline runs, even when the inputs and code haven't changed. They want to leverage pipeline caching to avoid redundant executions. What should they do to enable caching for this component?

Accepted Answer

Ensure that the component does not have 'dsl.cache_options(enable_cache=False)' set.. Vertex AI pipeline caching is enabled by default for all components unless explicitly disabled using `dsl.cache_options(enable_cache=False)`. The component re-executing every time indicates that caching was likely disabled on that specific component. Removing or ensuring this setting is not present will allow the pipeline to reuse cached outputs when inputs and code have not changed.

Answer

Set the 'caching' flag to 'True' in the pipeline definition using 'pipeline.caching = True'.

Answer

Set the environment variable 'ENABLE_CACHE' to 'true' on the pipeline run request.

Answer

Re-compile the pipeline with the '--enable-cache' flag.