Generative AI Leader Practice Test 42 — 15 Questions

Question 1

A global retailer uses a generative AI model to personalize product recommendations. They need to ensure that customer prompts and responses are not logged for model improvement to meet GDPR data minimization principles. Which configuration should they apply?

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

Accepted Answer

Disable prompt/response logging in Vertex AI endpoint settings. To comply with GDPR data minimization, the retailer should disable prompt/response logging. Vertex AI offers settings to control whether prompts and responses are stored for model improvement.

Answer

Use a third-party logging service with data deletion policies

Answer

Enable data logging for six months and then automatically delete

Answer

Anonymize all customer data before logging

Question 2

Which Google AI research organization is responsible for AlphaFold, a breakthrough in protein structure prediction?

Accepted Answer

Google DeepMind. Google DeepMind is the research lab behind AlphaFold, AlphaCode, and other notable achievements.

Answer

Google Research

Answer

Google Brain

Answer

X Development

Question 3

A data science team wants to decide between using a pre-built API (e.g., Vertex AI Gemini API) and fine-tuning a custom model for a specific business task. Which TWO factors are most important in making this build versus buy decision?

Accepted Answer

Cost of inference per query for pre-built API vs fine-tuned model. Option D is correct because the build-versus-buy decision hinges on the total cost of ownership, and inference cost per query directly determines whether paying per-token/per-request for a pre-built API like Vertex AI Gemini is cheaper than hosting and serving a fine-tuned model at scale. Option E is correct because fine-tuning a custom model requires a sufficient volume of high-quality, task-specific labeled data; if such data is unavailable, the pre-built API is the practical choice, making data availability a decisive factor. The unmarked options do not belong: A (UI color scheme) is a cosmetic design choice unrelated to model economics or feasibility, B (developer language preference) is a minor implementation detail that does not determine build-versus-buy value, and C (number of users) is a demand metric that only indirectly matters through its effect on inference cost and data volume, which are already captured by D and E.

Answer

Color scheme of the user interface

Answer

Developer preference for programming languages

Answer

Number of users who will interact with the system

Question 4

A global bank wants to deploy a generative AI assistant for employees across multiple European countries, each with strict data residency laws. Which deployment strategy is most compliant?

Accepted Answer

Deploy separate model instances in each country's cloud region.. Deploying separate model instances in each country's cloud region ensures that data never crosses national borders, directly complying with strict data residency laws like the GDPR's data localization requirements. This strategy uses regional cloud infrastructure (e.g., AWS eu-central-1, Azure westeurope) to keep both training and inference data within the specific jurisdiction, avoiding any cross-border data transfer.

Answer

Use a federated learning approach where data stays on-premises.

Answer

Deploy a single model in a US region and use data masking.

Answer

Use a third-party API that processes data outside Europe.

Question 5

Which TWO techniques can help improve the factual accuracy of a language model's outputs? (Choose two.)

Accepted Answer

Fine-tune on a domain-specific curated dataset.. Fine-tuning on a domain-specific curated dataset (C) directly adjusts the model's weights using high-quality, verified examples, teaching it to produce factually correct outputs for that domain. This reduces hallucinations by grounding the model in accurate, relevant data rather than relying solely on its pre-training distribution.

Answer

Decrease the max output tokens.

Answer

Increase the temperature parameter.

Answer

Use top-k random sampling.

Question 6

A developer needs to use the Vertex AI PaLM API to generate text embeddings for a large corpus of documents. Which model should they use?

Accepted Answer

textembedding-gecko@001. `textembedding-gecko@001` is the specific Vertex AI model designed for generating text embeddings, which convert text into dense vector representations. This model is optimized for semantic similarity, clustering, and retrieval tasks, making it ideal for processing a large corpus of documents. The other models are designed for code generation, text generation, or chat, not embeddings.

Answer

codey-bison@001

Answer

text-bison@001

Answer

chat-bison@001

Question 7

A developer is using the Vertex AI PaLM API and receives a 429 Resource Exhausted error. What is the most likely cause?

Accepted Answer

The user has exceeded the allowed number of requests per minute. HTTP 429 Resource Exhausted is the standard error returned when a client exceeds the rate limits or quota assigned to the API — most commonly the requests-per-minute (RPM) limit. In the Vertex AI PaLM API, each project has per-minute quotas for prediction requests, and exceeding them triggers this error. The correct remediation is to implement exponential backoff and/or request a quota increase.

Answer

The request payload is too large

Answer

The model is not available in the current region

Answer

The API key is invalid

Question 8

A company is piloting a GenAI feature for email drafting in Gmail. They want to measure productivity improvement. Which metric is MOST directly tied to the business goal of reducing time spent on email composition?

Accepted Answer

Reduction in average time-to-send per email. The primary business goal is to reduce the time employees spend composing emails. Measuring the average time-to-send per email directly quantifies this efficiency gain, as it captures the end-to-end duration from initiation to dispatch, which the GenAI feature aims to shorten by generating draft content.

Answer

Adoption rate of the GenAI feature among the pilot group

Answer

Increase in employee satisfaction survey scores

Answer

Reduction in total tokens consumed per email

Question 9

A company is using Vertex AI Model Registry to manage multiple versions of its custom generative model. They want to automatically route a percentage of traffic to a new model version for testing. What should they do?

Accepted Answer

Deploy both versions to the same endpoint and adjust traffic split settings. Vertex AI Endpoints support traffic splitting between model versions.

Answer

Set up a Cloud Tasks queue to distribute requests

Answer

Create a new endpoint for each version

Answer

Use a load balancer in front of the endpoints

Question 10

A startup wants to quickly prototype a conversational AI application using Gemini. They need free access during development and do not require VPC controls. Which access tier should they choose?

Accepted Answer

Google AI Studio free tier. Google AI Studio's free tier provides free access to Gemini models for rapid prototyping without requiring VPC controls or any billing setup. This aligns directly with the startup's need for quick, cost-free development iteration before moving to production.

Answer

Vertex AI on a pay-as-you-go basis

Answer

Vertex AI with VPC-SC

Answer

Gemini API via Cloud Run

Question 11

A company is piloting a GenAI code review assistant. Developers report that the assistant often suggests incorrect or insecure code snippets. The team wants to improve the assistant's reliability before expanding the pilot. Which approach should they prioritize?

Accepted Answer

Fine-tune the base model on a curated dataset of secure code review examples. Fine-tuning on a curated dataset of secure code review examples directly addresses the assistant's tendency to suggest incorrect or insecure code by adapting the model's behavior to the specific patterns and standards of secure coding. Unlike retrieval or parameter adjustments, this approach modifies the model's weights to prioritize security and correctness in its outputs, making it the most effective method for improving reliability in a targeted domain.

Answer

Use RAG to retrieve code snippets from a database of known vulnerabilities

Answer

Increase the model's temperature parameter to generate more diverse suggestions

Answer

Switch to a larger foundation model without additional tuning

Question 12

A company is building a customer service chatbot using Vertex AI Agent Builder. The chatbot needs to answer questions based on a large internal knowledge base stored in a Cloud Storage bucket. The team wants to ensure the model can reference the latest documents without fine-tuning. Which configuration should they use?

Accepted Answer

Ground the model with a Vertex AI Search data store connected to the Cloud Storage bucket. Grounding with a Vertex AI Search data store is the correct pattern for retrieval-augmented generation (RAG) in Vertex AI Agent Builder. The data store indexes the Cloud Storage documents and the agent retrieves relevant chunks at query time, so the model can cite the latest content without any fine-tuning. This keeps the knowledge base fresh — updating a document in the bucket automatically flows into subsequent answers.

Answer

Fine-tune a model on the knowledge base documents

Answer

Use a pre-built model with no additional configuration

Answer

Store the documents in BigQuery and use a BigQuery connector

Question 13

A developer needs to generate embeddings for text data to be used in a semantic search application. Which Google Cloud service should they use?

Accepted Answer

Vertex AI Embeddings API. Vertex AI Embeddings API is the correct choice because it provides a managed service to generate vector embeddings from text data, which are essential for semantic search applications that rely on understanding meaning rather than exact keyword matches. This API leverages large language models to convert text into high-dimensional vectors, enabling efficient similarity search using vector databases or nearest neighbor algorithms.

Answer

Document AI

Answer

Cloud Translation API

Answer

Cloud Speech-to-Text

Question 14

A startup is building a multimodal application that allows users to upload a photo of a plant and ask questions about its care. They want to use Google Cloud generative AI services. Which combination of services is MOST suitable?

Accepted Answer

Use Gemini on Vertex AI with a prompt that includes the image and the user’s question. Gemini on Vertex AI natively supports multimodal inputs, allowing the model to directly process the uploaded plant image alongside the user's care question in a single prompt. This eliminates the need for intermediate text extraction or image description generation, providing the most accurate and context-aware response for the application's requirements.

Answer

Use Cloud Vision API to extract text from the image, then send that text to a text‑only model

Answer

Use Vertex AI Agent Builder with a text‑only model and ignore the image

Answer

Use Imagen to generate a description of the plant, then feed that description to a text model

Question 15

A company wants to estimate the total cost of ownership (TCO) for a gen AI solution on Google Cloud. Which factors are most important?

Accepted Answer

Compute, storage, and API call costs. The total cost of ownership (TCO) for a generative AI solution on Google Cloud encompasses all operational expenses, including compute (e.g., TPU/GPU instances for training and inference), storage (e.g., Cloud Storage for datasets and model artifacts), and API call costs (e.g., Vertex AI prediction requests). Focusing on a single cost component, such as training or inference alone, ignores the recurring expenses of serving the model and storing data, which often dominate long-term TCO.

Answer

Only model training cost

Answer

Only inference cost

Answer

Only compute cost