Generative AI Leader Practice Test 30 — 15 Questions

Question 1

A data analyst wants to train and deploy a custom image classification model with minimal ML engineering overhead. Which Google Cloud service should they use?

Accepted Answer

Vertex AI. Vertex AI is the correct choice because it provides a fully managed, end-to-end platform for training and deploying custom machine learning models, including image classification, with minimal ML engineering overhead. It offers AutoML capabilities for automated model training, managed pipelines, and one-click deployment to a scalable endpoint, eliminating the need for infrastructure management.

Answer

Google Colab

Answer

AI APIs (Vision AI)

Answer

BigQuery ML

Question 2

A generative AI model for code generation sometimes produces syntactically incorrect code. The team wants to reduce syntax errors without retraining the entire model. Which approach is most effective?

Accepted Answer

Implement constrained decoding with grammar rules. Constrained decoding with grammar rules directly enforces the syntax of the target programming language during token generation, preventing the model from producing invalid constructs. This approach modifies the decoding process (e.g., using a context-free grammar or a formal syntax specification) to mask or forbid tokens that would lead to a syntax error, without altering the underlying model weights. It is the most effective method because it guarantees syntactically correct output at generation time, rather than relying on post-hoc fixes or probabilistic adjustments.

Answer

Run a syntax checker after generation and regenerate

Answer

Add a system prompt that instructs the model to produce valid code

Answer

Increase beam search width

Question 3

A company is fine-tuning a generative AI model on proprietary customer data. They are concerned about copyright and IP issues when using the model commercially. What is the BEST practice to mitigate these risks?

Accepted Answer

Only use data that is explicitly licensed for commercial use and document its provenance. The most reliable way to mitigate copyright and IP risk when fine-tuning on proprietary data is to ensure every piece of training data is explicitly licensed for commercial use and to maintain clear documentation of its provenance. This creates an auditable chain of custody that supports both legal defensibility and compliance. It addresses the root cause—data rights—rather than trying to patch symptoms after generation.

Answer

Apply SynthID to all generated content to prove origin

Answer

Include a disclaimer on all outputs that the company is not liable for IP infringement

Answer

Use a model trained on publicly available data only

Question 4

A retail company is building a product description generator using a large language model on Vertex AI. They need to ensure the generated descriptions do not contain offensive language. Which strategy should they implement?

Accepted Answer

Implement a content moderation filter (e.g., Perspective API) as a post-processing step. Content moderation filters like Perspective API act as a post-processing safeguard that can catch offensive language the model might generate despite prompt engineering or fine-tuning. This approach provides a deterministic, rule-based or ML-based check that is independent of the model's training, ensuring compliance with content policies in production. It is a standard practice for deploying LLMs in customer-facing applications where safety is critical.

Answer

Fine-tune the model on a dataset of clean product descriptions

Answer

Use Vertex AI Model Monitoring to detect anomalies in model predictions

Answer

Include explicit instructions in the prompt to avoid offensive language

Question 5

A healthcare startup has fine-tuned a Vertex AI PaLM 2 model on a dataset of medical records to generate patient summaries. The model produces fluent text but occasionally fabricates diagnoses not present in the input. The team has already tried increasing the training data size by 20% and adjusting the temperature from 0.7 to 0.2, but hallucinations persist. The summaries must be factually accurate for regulatory compliance. What should the team do next?

Accepted Answer

Implement a RAG pipeline using Vertex AI Search to retrieve relevant medical documents before generation.. Implementing a Retrieval-Augmented Generation (RAG) pipeline with Vertex AI Search grounds the model's output in retrieved, authoritative medical documents. This directly addresses the root cause of hallucination—the model's reliance on its parametric memory—by providing factual context at inference time, which is far more effective for regulatory compliance than adjusting generation parameters or training data size alone.

Answer

Increase the maximum output tokens to allow the model to generate more detailed summaries.

Answer

Add more few-shot examples to the prompt for each generation.

Answer

Switch the base model to Gemini 1.5 Pro without additional changes.

Question 6

Which THREE considerations are critical when deploying a generative AI model using Vertex AI Endpoints for a latency-sensitive application? (Choose THREE.)

Accepted Answer

Model size and architecture. Model size and architecture directly impact inference latency because larger models with more parameters require more computation per request. For latency-sensitive applications, choosing a smaller or distilled model (e.g., Gemma 2B vs. 27B) or using quantization can reduce response times. Vertex AI Endpoints serve the model as-is, so the model's inherent computational cost is the primary driver of per-request latency.

Answer

Number of model versions

Answer

Number of model instances

Question 7

A healthcare chatbot must avoid hallucinations. Which TWO techniques should the team implement? (Choose two.)

Accepted Answer

Use chain-of-thought prompting. Chain-of-thought prompting (B) reduces hallucinations by forcing the model to reason step-by-step, which improves factual accuracy and consistency in complex tasks like medical triage. Enabling grounding with a knowledge base (E) anchors the model's output to verified external data, directly preventing fabrication by restricting responses to retrieved facts.

Answer

Set frequency penalty to 0.0

Answer

Use higher temperature

Answer

Increase top_k to 50

Question 8

A company is deciding between using a pre-built GenAI API (like Gemini API) and building a custom fine-tuned model. Which factor would MOST strongly favor the custom fine-tuned model?

Accepted Answer

Need for specialized domain knowledge that general models lack. A custom fine-tuned model is most justified when the task requires deep, specialized domain knowledge that general-purpose models cannot reliably provide. Fine-tuning adapts a base model's weights to a narrow corpus (e.g., proprietary medical, legal, or industrial data), improving accuracy and style adherence beyond what prompt engineering alone can achieve. The other factors—budget, latency, and request volume—are operational constraints that typically favor pre-built APIs, not custom models.

Answer

Limited budget for AI development

Answer

Low latency requirements for real-time responses

Answer

Small volume of inference requests per day

Question 9

A marketing team wants to use GenAI to generate social media posts. They need consistent brand voice and the ability to output structured JSON for downstream scheduling. Which combination of Vertex AI features should they use?

Accepted Answer

Vertex AI Studio with few-shot prompting and a response schema for JSON. Vertex AI Studio offers prompt design capabilities, and you can define few-shot examples to enforce tone. Structured output can be requested by specifying a response schema. Model Garden is a model hub, not a prompt tool; Fine-tuning is overkill for this task.

Answer

Model Garden to select a model and then fine-tune it for brand voice

Answer

Vertex AI RAG Engine with a vector store of past posts

Answer

Vertex AI Agent Builder and Grounding with Google Search

Question 10

A financial institution wants to ensure compliance with GDPR when using a generative AI service that processes EU user data. Which measure is most directly required?

Accepted Answer

Implement a mechanism to obtain user consent before processing data. GDPR requires a lawful basis for processing personal data, and consent is one of the most common bases, especially for marketing or non-contractual processing. Implementing a mechanism to obtain user consent before processing EU user data directly satisfies GDPR Article 6 and Article 7 requirements for lawful processing. This is the most directly required measure among the options.

Answer

Disable logging of all interactions to minimize data retention

Answer

Store all prompts and responses in a US-based data center

Answer

Use a model trained only on non-EU data

Question 11

A healthcare startup wants to use generative AI to provide clinical decision support. They must minimize the risk of harmful hallucinations. Which business strategy is most appropriate?

Accepted Answer

Implement retrieval-augmented generation with meticulously curated medical literature.. Retrieval-augmented generation (RAG) grounds the model's output in a trusted, external knowledge base—here, curated medical literature—which directly reduces the risk of hallucination by forcing the model to cite or derive answers from verified sources. This is the most effective strategy for clinical decision support because it combines generative flexibility with factual accuracy, unlike methods that only limit output or rely on post-hoc filtering.

Answer

Limit the model's output length to reduce hallucination risk.

Answer

Deploy a large general-purpose model and rely on post-processing filters.

Answer

Use a custom fine-tuned model on a proprietary medical dataset.

Question 12

Refer to the exhibit. The endpoint is experiencing high latency during traffic spikes. The team wants to improve response time by reducing queueing. Which change to the configuration would be most effective?

Accepted Answer

Increase maxReplicaCount to 10. Increasing maxReplicaCount to 10 allows the autoscaler to provision more replicas during traffic spikes, distributing the incoming requests across additional endpoints. This directly reduces queueing at each replica because the load is spread over more instances, lowering per-instance latency. The change targets the root cause—insufficient capacity to handle peak load—rather than adjusting thresholds or model versions.

Answer

Decrease minReplicaCount to 0

Answer

Change the model version to '2'

Answer

Decrease the target value in autoscaling metric to 50

Question 13

A company is building a customer support chatbot using Vertex AI Agent Builder. They want the agent to answer questions based on internal knowledge base documents stored in Cloud Storage. Which feature should they configure to ensure the agent can retrieve relevant information from these documents?

Accepted Answer

Enable grounding with a data store. Vertex AI Agent Builder uses grounding to connect the agent to external data sources, such as documents stored in Cloud Storage. By enabling grounding with a data store, the agent can retrieve and reference relevant information from the knowledge base documents in real time, ensuring accurate and context-aware responses without requiring model retraining.

Answer

Deploy the agent to a Vertex AI endpoint

Answer

Fine-tune a Gemini model on the knowledge base

Answer

Configure a safety filter to block irrelevant queries

Question 14

Which THREE factors should be considered when choosing between Gemini 1.5 Pro and Gemini 1.5 Flash for a customer-facing chatbot? (Choose three.)

Accepted Answer

Cost constraints: Flash is more cost-effective per token. Gemini 1.5 Flash is designed as a cost-optimized model, offering significantly lower per-token pricing compared to Gemini 1.5 Pro. For a customer-facing chatbot with high query volumes, cost efficiency is a primary consideration, making Flash the more economical choice for routine interactions.

Answer

Safety filters: Pro has stricter safety defaults

Answer

Multimodal capability: Flash does not support image input

Question 15

What is the primary purpose of the temperature parameter in a generative language model?

Accepted Answer

It controls the tradeoff between creativity and determinism. The temperature parameter directly controls the probability distribution over the next token. A lower temperature (e.g., 0.1) makes the model more deterministic by favoring high-probability tokens, while a higher temperature (e.g., 1.5) flattens the distribution, increasing randomness and creative output. This tradeoff is fundamental to balancing coherence with novelty in generative models.

Answer

It defines the context window size for the prompt

Answer

It limits the vocabulary to the top K tokens

Answer

It sets the maximum number of tokens in the output