Generative AI Leader Practice Test 19 — 15 Questions

Question 1

A data scientist wants to quickly prototype a text generation application using Google's foundation models. Which Google Cloud service should they use?

Accepted Answer

Generative AI Studio. Generative AI Studio is the correct service because it provides a purpose-built environment for quickly prototyping and experimenting with Google's foundation models, including text generation models like PaLM 2 and Gemini. It offers a no-code interface and SDK access for rapid iteration, directly aligning with the data scientist's goal of fast prototyping without needing to manage infrastructure or training pipelines.

Answer

Cloud Natural Language API

Answer

Vertex AI Prediction

Answer

AI Platform Training

Question 2

A team wants to build a GenAI application that can interact with external APIs (e.g., to check inventory or place orders). Which Vertex AI component provides this capability?

Accepted Answer

Vertex AI Extensions. Vertex AI Extensions allows a GenAI application to connect to and interact with external APIs, enabling actions like checking inventory or placing orders. It provides a way to extend the model's capabilities by calling external services. This is the component specifically designed for such integrations.

Answer

Grounding with Google Search

Answer

Model Garden

Answer

Vertex AI Agent Builder

Question 3

A project manager wants to reduce the cost of using Gemini API for batch processing of customer feedback. The team is on a tight budget. Which cost management strategy is MOST effective?

Accepted Answer

Use batch requests to group multiple prompts. Batch requests reduce per-token cost by grouping multiple prompts into one API call. Caching helps but is less impactful for varied feedback.

Answer

Increase the temperature to max to reduce output length

Answer

Disable logging to reduce storage costs

Answer

Switch to the largest model available for better accuracy

Question 4

A media company is using Vertex AI Imagen to generate marketing images. The output frequently contains unrealistic artifacts, especially in human faces. The team has fine-tuned the model using their brand assets. What is the most likely cause and recommended fix?

Accepted Answer

The fine-tuning dataset is too small or too homogeneous; augment and diversify the training data.. Unrealistic artifacts in fine-tuned generative models, especially in human faces, typically stem from a training dataset that is too small or lacks diversity. When the dataset is homogeneous, the model overfits to limited patterns and fails to generalize, leading to distorted outputs. Augmenting and diversifying the training data with varied poses, lighting, and ethnicities helps the model learn robust facial features.

Answer

Safety filters are too aggressive; reduce them.

Answer

Negative prompts are missing; always include 'unrealistic'.

Answer

Inference steps are too low; increase to 100.

Question 5

Which THREE approaches are effective for reducing bias in generative model outputs? (Choose three.)

Accepted Answer

Use a balanced training dataset.. A balanced training dataset reduces the risk of the model learning spurious correlations or skewed distributions that lead to biased outputs. By ensuring that all demographic groups, topics, or perspectives are represented proportionally, the model's learned probability distribution is less likely to favor one group over another, directly mitigating representation bias at the data level.

Answer

Set temperature to a very high value.

Answer

Use adversarial training.

Question 6

Refer to the exhibit. A Vertex AI endpoint configured with the above deployment is returning HTTP 429 (Too Many Requests) errors during peak traffic. The current CPU utilization reaches 80% consistently. What should the team adjust to resolve this?

Accepted Answer

Increase maxReplicaCount to 10. Increasing maxReplicaCount to 10 allows the Vertex AI endpoint to scale out to more instances during peak traffic, distributing the load and reducing HTTP 429 errors. Since CPU utilization is at 80%, the current maxReplicaCount is insufficient to handle the demand, and raising this limit enables the horizontal pod autoscaler to add replicas up to the new maximum, directly addressing the capacity bottleneck.

Answer

Increase scaleTarget to 0.9

Answer

Change machineType to n1-highmem-2

Answer

Increase minReplicaCount to 2

Question 7

A multinational corporation deploys a generative AI chatbot for customer support in the EU. They must ensure compliance with GDPR regarding user data used for fine-tuning. Which data governance practice is REQUIRED?

Accepted Answer

Obtain explicit consent from users for their data to be used in fine-tuning. Under GDPR, using personal data for fine-tuning a generative AI model is a new processing purpose and requires its own lawful basis. Where consent is used, it must be freely given, specific, informed and unambiguous, and explicit consent is required for special-category or unexpected secondary uses. Option B identifies the mandatory legal requirement: obtaining explicit consent from users for their data to be used in fine-tuning. Option C addresses a data subject right that may apply once data is already processed, but it is not the primary upfront governance practice required for using personal data in fine-tuning.

Answer

Store user data only in the EU region to comply with data residency requirements

Answer

Implement a mechanism to delete specific user data from the fine-tuning dataset upon user request

Answer

Anonymize all user data before using it for fine-tuning

Question 8

A healthcare company is using a generative AI model to summarize patient notes. The model occasionally includes hallucinated medical details. Which strategy best reduces hallucinations?

Accepted Answer

Provide the model with the patient notes as context and instruct it to only use that information.. Grounding the model by providing the patient notes as context and instructing it to only use that information is the most effective way to reduce hallucinations. This approach, often called retrieval-augmented generation or prompt grounding, forces the model to base its summary on the given text rather than relying on its parametric memory. It also allows for verification against the source.

Answer

Use a larger model with more parameters.

Answer

Fine-tune the model on a large corpus of general medical literature.

Answer

Increase the model's temperature to encourage more diverse outputs.

Question 9

Refer to the exhibit. A data scientist is fine-tuning a model. The training loss and accuracy are improving each epoch. However, after training, the model performs poorly on a held-out validation set. What is the most likely issue?

Accepted Answer

Overfitting. The model's training loss and accuracy improve each epoch, but performance on the validation set is poor. This classic symptom indicates overfitting, where the model memorizes the training data (including noise) rather than learning generalizable patterns. In fine-tuning, this often occurs when the model is trained for too many epochs or the dataset is too small relative to model capacity.

Answer

Underfitting

Answer

Inappropriate learning rate

Answer

Data leakage

Question 10

A developer is using Vertex AI Studio to experiment with prompts. They want to ensure that the model's responses are grounded in factual information from a trusted knowledge base. Which feature should they enable?

Accepted Answer

Grounding with a Vertex AI Search data store. Vertex AI's grounding feature allows the model to cite sources from a provided knowledge base, improving factual accuracy and verifiability.

Answer

Safety filters

Answer

Temperature setting reduction

Answer

Chain-of-thought prompting

Question 11

A company is migrating an on-premises NLP pipeline to Vertex AI. Which three capabilities of Vertex AI align with common MLOps best practices for generative AI? (Choose THREE)

Accepted Answer

Automatic model retraining based on performance degradation. Option A is correct because Vertex AI Model Monitoring can detect training-serving skew and prediction drift (e.g., via drift/skew thresholds on feature distributions) and trigger automated retraining workflows, which is a core MLOps practice for keeping generative AI models performant over time. Option C is correct because Vertex AI Pipelines (built on Kubeflow Pipelines/Vertex AI Pipelines SDK) enables continuous training by orchestrating reproducible DAGs that ingest new data, retrain, evaluate, and redeploy models as part of a CI/CD/CT pipeline. Option E is correct because Vertex AI Model Registry provides centralized versioning, lineage, and metadata tracking of model artifacts, allowing teams to manage model versions, roll back, and promote models to endpoints—essential for governed MLOps. Option B is not correct because Vertex AI is a managed cloud service on Google Cloud, not a local on-premises execution environment. Option D is not correct because Vertex AI supports automated and programmatic labeling (e.g., Vertex AI Data Labeling Service with human-in-the-loop and active learning), so restricting to manual labeling only contradicts MLOps best practices.

Answer

Local on-premises execution

Answer

Manual data labeling only

Question 12

Refer to the exhibit. A sudden surge of traffic reaches 15,000 requests per second, but the endpoint can only handle 1,000 req/s per replica. What will happen to new requests?

Accepted Answer

They will receive HTTP 429 (Too Many Requests) errors.. When a surge of 15,000 requests per second hits an endpoint configured with a maxReplicaCount (e.g., 10 replicas at 1,000 req/s each = 10,000 req/s capacity), any excess requests beyond that capacity are rejected with an HTTP 429 (Too Many Requests) status code. This is standard behavior in autoscaling systems: once the replica count reaches its maximum limit, the service cannot scale further, and new requests are throttled to prevent overload.

Answer

They will be processed, and replicas will exceed maxReplicaCount.

Answer

They will be redirected to a different model.

Answer

They will be queued until capacity becomes available.

Question 13

A developer wants to add a GenAI feature to their existing web application. They need to integrate with the app's backend using REST APIs. Which integration pattern is MOST appropriate?

Accepted Answer

Integrate via Vertex AI API or Gemini API. API-first integration using Vertex AI API or Gemini API is the standard way to add GenAI to existing applications. Workspace add-ons are for Google Workspace apps. Apps Script is for automating Workspace. Agent Builder is for building conversational agents, not general API integration.

Answer

Use Apps Script to call the Gemini API

Answer

Use Vertex AI Agent Builder to create an agent and embed it via iframe

Answer

Build a Google Workspace add-on

Question 14

A company is evaluating ROI for a GenAI-based code review assistant. Which metric set BEST captures both productivity and quality improvements?

Accepted Answer

Time saved per code review and bug detection rate (percentage of bugs caught before deployment). Time saved per review (productivity) and bug detection rate (quality) directly measure the tool's impact. Code churn and developer satisfaction are secondary. Defect escape rate is important but harder to measure directly for code review.

Answer

Number of reviews completed per day and lines of code written

Answer

Model inference latency and cost per token

Answer

Developer satisfaction score and reduction in code churn (percentage of code rewritten)

Question 15

An enterprise is concerned about the cost of using a large LLM for a high-volume customer support chatbot. They want to reduce token consumption while maintaining response quality. Which strategy would be MOST effective?

Accepted Answer

Implement response caching for common questions and batch similar requests. Caching frequent queries reduces costs because the response is served from cache without model inference. Batching requests also saves on per-request overhead. Choosing a smaller model may reduce quality.

Answer

Use a multimodal model to handle both text and images

Answer

Always use the largest available model to ensure best quality

Answer

Increase the max output tokens to capture more detail