Generative AI Leader Generative AI Concepts and Technologies Practice Test 8 — 15 Questions

Question 1

An enterprise is deploying a generative AI solution that must comply with GDPR data residency requirements. They plan to use Vertex AI with Gemini. Which configuration is necessary?

Accepted Answer

Deploy Gemini on Vertex AI with VPC Service Controls and Customer-Managed Encryption Keys (CMEK). Option D is correct because GDPR data residency requires that data remain within a specific geographic boundary. VPC Service Controls provide a security perimeter that prevents data exfiltration from the chosen region, while CMEK ensures encryption keys are managed by the customer, not Google, meeting the control requirements for data residency compliance.

Answer

Use the default settings on Vertex AI with Gemini, as Google automatically handles data residency

Answer

Fine‑tune Gemini on a private Google Cloud Storage bucket in the target region

Answer

Use the Gemini API directly with a restricted API key

Question 2

What is the key advantage of using adapter-based fine-tuning methods like LoRA compared to full fine-tuning of a large language model?

Accepted Answer

LoRA significantly reduces the number of trainable parameters, making fine-tuning more memory-efficient. LoRA (Low-Rank Adaptation) injects trainable low-rank matrices into the transformer layers while keeping the original model weights frozen. This drastically reduces the number of trainable parameters (often by 10,000x), which lowers GPU memory requirements for storing optimizer states and gradients during training, making fine-tuning feasible on consumer hardware.

Answer

LoRA is faster at inference time compared to the fully fine-tuned model

Answer

LoRA eliminates the need for a base model

Answer

LoRA enables training on a larger dataset than full fine-tuning

Question 3

A developer is building a code‑generation assistant using the Codey API on Vertex AI. The assistant should generate Python functions based on natural language descriptions. However, the generated code sometimes contains syntax errors. Which parameter adjustment would MOST directly help reduce syntax errors?

Accepted Answer

Lower the temperature (e.g., from 0.8 to 0.2). Reducing temperature makes the model more deterministic, which typically reduces creative but incorrect outputs like syntax errors. Prompt engineering can also help, but adjusting temperature is the simplest direct fix. Increasing max tokens or changing top-k does not directly address syntax correctness.

Answer

Increase the context window

Answer

Set top-k to 1

Answer

Increase the max output tokens

Question 4

A data scientist is using Vertex AI to fine‑tune a PaLM 2 model for a legal document summarization task. They have 10,000 labeled document‑summary pairs. After supervised fine‑tuning, the model performs well on the training set but often hallucinates names and dates on unseen documents. Which next step is MOST likely to improve factual accuracy?

Accepted Answer

Apply reinforcement learning from human feedback (RLHF) using a preference dataset that penalizes factual inaccuracies. RLHF (Reinforcement Learning from Human Feedback) is specifically designed to align model outputs with human preferences, which can reduce hallucinations by penalizing factually incorrect generations. More data or longer training may not fix the underlying alignment issue. RAG is a separate approach but RLHF directly addresses the hallucination from the model's behavior.

Answer

Increase the number of fine‑tuning epochs to 10

Answer

Use a larger base model like Gemini Ultra without fine‑tuning

Answer

Add more training examples from a different domain

Question 5

A financial institution wants to use Gemini to analyze customer support transcripts and generate summaries. They need to ensure that personally identifiable information (PII) is not included in the summaries. Which approach should they take?

Accepted Answer

Preprocess the transcripts with Cloud DLP API to redact PII before sending to Gemini. Option A is correct because the Cloud Data Loss Prevention (DLP) API provides a purpose-built, scalable service to detect and redact PII from text before it reaches the Gemini model. This ensures that sensitive data is removed at the source, preventing any possibility of leakage in the generated summary, regardless of model behavior or prompt engineering.

Answer

Use a carefully engineered prompt instructing Gemini not to include PII

Answer

Post‑process the generated summaries with a regex filter to remove PII

Answer

Fine‑tune Gemini to avoid generating PII

Question 6

Which statement best describes the difference between the Gemini Flash and Gemini Pro models on Vertex AI?

Accepted Answer

Gemini Flash is optimized for speed and cost, while Gemini Pro provides higher quality for complex tasks. Option C is correct because Gemini Flash is specifically designed for low-latency, high-throughput, and cost-efficient inference, making it ideal for high-volume, simpler tasks. In contrast, Gemini Pro is a larger, more capable model that delivers superior quality and reasoning for complex, multi-step tasks, though at higher latency and cost. This distinction is fundamental to the Gemini model family on Vertex AI, where Flash serves as the lightweight, fast option and Pro as the premium, high-quality option.

Answer

Gemini Flash is a distilled version of Gemini Pro that requires fine‑tuning before use

Answer

Gemini Pro is deployed on Google’s TPU v5p chips, while Flash uses TPU v4

Answer

Gemini Flash is only available for image inputs, while Gemini Pro handles text

Question 7

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. Option C is correct because 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 8

A company is deploying a generative AI application using Vertex AI. They need to minimize latency for real‑time inference while maintaining high quality. Which TWO actions are most effective?

Accepted Answer

Use a smaller model like Gemini Flash. Using a smaller model (e.g., Gemini Flash) reduces latency, and reducing the max output tokens limits generation time. Increasing temperature does not affect latency; using a larger model increases latency; batching is for throughput, not single‑request latency.

Answer

Batch multiple inference requests together

Answer

Use Gemini Pro instead of Gemini Flash to ensure quality

Answer

Increase the temperature to 1.0 for more creative outputs

Question 9

A data scientist is evaluating how to ground a generative AI model to reduce hallucinations when answering questions about a private knowledge base. Which TWO techniques are most suitable?

Accepted Answer

Retrieval-Augmented Generation (RAG). Retrieval-Augmented Generation (RAG) is the most suitable technique because it retrieves relevant, up-to-date chunks from the private knowledge base at inference time and conditions the generative model's output on that retrieved context. This grounds the model in factual data, directly reducing hallucinations by ensuring answers are based on the retrieved evidence rather than the model's parametric memory alone.

Answer

Using a larger model like Gemini Ultra

Answer

Fine‑tuning on the private knowledge base

Answer

Increasing the temperature to 0.9

Question 10

Which THREE of the following are generative AI modalities supported by Google Cloud services?

Accepted Answer

Image generation. Option A is correct because Google Cloud's Vertex AI and Imagen APIs provide image generation capabilities, allowing users to create and edit images from text prompts. This is a core generative AI modality supported by Google Cloud services.

Answer

Tabular data generation

Answer

Speech generation

Question 11

A company is fine‑tuning a large language model for a domain‑specific task. They have a limited budget and want to minimize the cost of fine‑tuning. Which TWO approaches are most cost‑effective?

Accepted Answer

Use a smaller base model like Gemini Flash. Option A is correct because using a smaller base model like Gemini Flash reduces the computational resources (GPU memory, training time) required for fine-tuning, directly lowering cost while still achieving adequate performance for domain-specific tasks. Smaller models have fewer parameters, which means fewer floating-point operations (FLOPs) per training step, making them more budget-friendly.

Answer

Use full fine‑tuning for better quality

Answer

Increase the number of training epochs

Answer

Use a larger base model like Gemini Ultra

Question 12

An organization is building a multi‑agent workflow on Vertex AI where one agent analyzes an image (e.g., a scanned contract), another agent extracts text from the image, and a third agent answers questions about the contract. The solution must be low‑latency. Which THREE services are most appropriate?

Accepted Answer

Gemini on Vertex AI. Gemini on Vertex AI (C) is the most appropriate service because it is a multimodal model that can natively analyze images, extract text, and answer questions about the content in a single, low-latency inference call. This eliminates the need to chain separate services for image analysis, OCR, and Q&A, reducing overall latency and architectural complexity.

Answer

Imagen

Answer

Vertex AI Search

Answer

Document AI

Question 13

A company wants to build a customer service chatbot that answers questions about their internal policy documents. The documents are updated monthly, and the team cannot afford to retrain a model each time. Which approach is MOST appropriate?

Accepted Answer

Use Retrieval-Augmented Generation (RAG) with the policy documents indexed in a vector store. RAG (Retrieval-Augmented Generation) allows the LLM to retrieve relevant document sections at inference time, so knowledge stays current without retraining. The other options either require expensive retraining for each update or lack document grounding.

Answer

Use a larger foundation model with a longer context window and paste all documents into each prompt

Answer

Train a custom model from scratch on the policy documents each month

Answer

Fine-tune a base LLM on the policy documents monthly

Question 14

A data scientist is fine-tuning a large language model for a legal document summarization task. The dataset contains only 500 examples, and the model must not forget its general language capabilities. Which fine-tuning method is most suitable?

Accepted Answer

Adapter-based fine-tuning using LoRA. LoRA (Low-Rank Adaptation) is an adapter-based fine-tuning method that updates a small number of parameters while keeping the base model frozen. This prevents catastrophic forgetting and works well with small datasets. Full fine-tuning would risk overfitting and losing general capabilities.

Answer

Retraining the model from scratch on the legal dataset

Answer

In-context learning with a few examples in the prompt

Answer

Full fine-tuning of all model parameters

Question 15

A media company wants to generate video content from text descriptions. They need a Google Cloud solution that can produce high-quality videos with realistic motion. Which TWO services should they consider?

Accepted Answer

Veo. Veo is Google Cloud's advanced video generation model that creates high-quality videos with realistic motion from text descriptions. It leverages diffusion transformers and temporal coherence techniques to ensure smooth, lifelike movement across frames, making it the correct choice for this use case.

Answer

Chirp

Answer

Codey

Answer

Gemini (multimodal)

Answer

Imagen