AIF-C01 Practice Test 36 — 15 Questions

Question 1

An organization needs to generate high-quality images from text prompts for a marketing campaign. They require the ability to edit specific regions of an image (inpainting) and extend images beyond their original boundaries (outpainting). Which AWS service or model should they choose?

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

Accepted Answer

Amazon Titan Image Generator. Amazon Titan Image Generator is the correct choice because it natively supports both inpainting (editing specific regions of an image) and outpainting (extending images beyond their original boundaries) through its image conditioning capabilities. This service is specifically designed for generative image tasks like text-to-image generation, inpainting, and outpainting, making it ideal for the marketing campaign's requirements.

Answer

Stable Diffusion XL via Amazon Bedrock

Answer

Amazon Rekognition

Answer

Amazon SageMaker JumpStart with a custom GAN

Question 2

A developer is using Amazon Titan Text Express via the Bedrock InvokeModel API to summarize long documents. The documents are approximately 6,000 tokens each, but the model's context window is 4,000 tokens. What is the BEST approach to handle this?

Accepted Answer

Split the document into chunks of 3,500 tokens, summarize each chunk, then combine the summaries. It uses a chunking-and-summarization strategy to handle documents that exceed the model's context window. Amazon Titan Text Express has a 4,000-token context limit, so splitting the 6,000-token document into smaller chunks (e.g., 3,500 tokens each), summarizing each chunk independently via the InvokeModel API, and then combining those summaries produces a final summary without exceeding the model's token limit. Option D is incorrect because although Amazon Titan Text Premier has a 32,000-token context window, switching models is not the best approach given the developer is already using Titan Text Express; the chunking strategy is a more direct and cost-effective solution that works with the existing model.

Answer

Truncate the document to 4,000 tokens by removing the middle portion

Answer

Use the Converse API instead, which supports longer context natively

Answer

Switch to Amazon Titan Text Premier, which has a larger context window

Question 3

A company is deploying a generative AI application on Amazon Bedrock that must meet strict latency requirements for real-time user interactions. Which THREE factors should they consider when selecting a foundation model? (Select THREE.)

Accepted Answer

The context window length. Option A (context window length) is correct because the number of tokens a model can ingest and generate directly affects per-request processing time and memory usage; larger context windows generally increase latency, so it must be evaluated against real-time requirements. Option B (the size of the model, i.e., parameter count) is correct because larger parameter counts typically require more compute per token, increasing inference latency and reducing tokens-per-second throughput, which is critical for real-time interactions. Option E (the model provider's regional infrastructure and supported throughput) is correct because latency depends on network proximity to the Bedrock endpoint, regional availability, and the provisioned or on-demand throughput/quota limits that determine whether the application can sustain real-time response times. Option C is not correct because training dataset size does not determine inference latency once the model is trained and served. Option D is not correct because the fine-tuning framework affects customization workflow and cost, not the runtime latency characteristics used to select a foundation model for real-time interactions.

Answer

The size of the model's training dataset

Answer

The fine-tuning framework used to customize the model

Question 4

A developer wants to invoke a foundation model in Amazon Bedrock with a large number of similar requests for cost savings. Which feature can help reduce inference cost?

Accepted Answer

Model caching. Model caching in Amazon Bedrock stores the results of frequently used or repeated inference requests, allowing the system to serve identical prompts from a cache rather than re-running the foundation model. This directly reduces inference cost by avoiding redundant compute for similar requests, making it the ideal choice for a developer with a large number of similar requests.

Answer

Fine-tuning

Answer

Prompt management

Answer

Batch inference

Question 5

A healthcare company needs to build a GenAI application that summarizes patient discharge notes. The compliance team requires that no medical record identifiers (MRIs) appear in the model's output. Which Bedrock Guardrails feature should be configured?

Accepted Answer

PII detection and redaction. Bedrock Guardrails' sensitive information filters (PII detection and redaction) can identify and redact standard PII such as names, addresses, and Social Security numbers. Medical record identifiers (MRIs) are not a built-in PII category, so you must add a custom regex pattern under the sensitive information filter to match and redact the MRI format. This is the correct feature because it prevents MRIs from appearing in output.

Answer

Topic denial

Answer

Grounding check

Answer

Content filtering

Question 6

A data scientist wants to quickly prototype different prompt variants for a summarization task using Amazon Bedrock. Which tool should they use for an interactive, no-code experimentation environment?

Accepted Answer

Amazon Bedrock Playground. Amazon Bedrock Playground provides an interactive, no-code web interface within the AWS Management Console where data scientists can quickly test and iterate on different prompt variants for foundation models. It supports real-time experimentation with models like Claude, Llama, and Titan without writing any code, making it ideal for rapid prototyping of summarization tasks.

Answer

Amazon SageMaker Canvas

Answer

Amazon Bedrock Studio

Answer

AWS Cloud9

Question 7

A data science team is evaluating a text classification model built using Amazon Bedrock. They want to compare its performance against a baseline using automated metrics. Which three metrics are appropriate for evaluating a text classification model? (Choose THREE.)

Accepted Answer

Recall. Recall (B) is correct because it measures the proportion of actual positive cases that the classifier correctly identified, which is a core metric for text classification performance. Precision (D) is correct because it measures the proportion of predicted positive cases that are truly positive, complementing recall for classification evaluation. Accuracy (E) is correct because it measures the overall fraction of correct predictions across all classes, making it a standard baseline metric for text classification. BLEU (A) is not appropriate because it is a precision-oriented n-gram overlap metric designed for machine translation and text generation, not classification. ROUGE (C) is not appropriate because it is a recall-oriented n-gram overlap metric used for summarization and generation tasks, not for evaluating classification labels.

Answer

BLEU

Answer

ROUGE

Question 8

A developer is using Amazon Bedrock Knowledge Bases to build a Q&A chatbot. The knowledge base contains PDF documents that are ingested and chunked. After ingestion, the chatbot sometimes returns irrelevant answers. What is the most likely cause?

Accepted Answer

The chunking size is too large, causing chunks to contain multiple topics. When chunking size is too large, each chunk may contain multiple distinct topics or concepts. The embedding model then generates a single vector for that chunk, which dilutes the semantic focus and causes the retriever to return chunks that are only partially relevant to the query, leading to irrelevant answers.

Answer

The embedding model is too large

Answer

The vector store index type is incorrect

Answer

The LLM temperature is set too low

Question 9

A developer is building an agent using Amazon Bedrock Agents. The agent needs to call an external API to retrieve weather data. What must the developer define to enable this capability?

Accepted Answer

An action group with an OpenAPI schema and a Lambda function. To enable an Amazon Bedrock Agent to call an external API, the developer must define an action group that includes an OpenAPI schema describing the API's operations and a Lambda function to execute the API call and return results. The action group bridges the agent's intent recognition with the actual API invocation, making option B the correct choice.

Answer

A Knowledge Base with weather documents

Answer

A Bedrock Guardrail to filter the weather data

Answer

A prompt flow in Bedrock Studio

Question 10

A data scientist is evaluating two different foundation models for a summarization task. They want to compare the quality of summaries generated by each model against a set of human-written reference summaries. Which set of metrics is most appropriate for this automated evaluation?

Accepted Answer

ROUGE, BLEU, and BERTScore. ROUGE, BLEU, and BERTScore are specifically designed for evaluating text generation quality against reference summaries. ROUGE measures n-gram overlap (recall-oriented), BLEU measures precision of n-gram matches, and BERTScore uses contextual embeddings from BERT to capture semantic similarity, making them ideal for summarization tasks.

Answer

Accuracy, precision, recall, and F1-score

Answer

Latency and throughput

Answer

Mean squared error (MSE) and R-squared

Question 11

A developer is using Amazon Bedrock to build a generative AI application. The application must deny any user request that involves asking about the topic of 'mergers and acquisitions'. Which Bedrock feature should the developer use?

Accepted Answer

Bedrock Topic Denial (part of Guardrails). Bedrock Guardrails includes a Topic Denial feature that allows developers to define specific topics (e.g., 'mergers and acquisitions') that the model must refuse to discuss. When a user input or model output matches a denied topic, Guardrails blocks the response and returns a predefined denial message, ensuring compliance with organizational policies.

Answer

Bedrock PII Detection

Answer

Bedrock Content Filtering

Answer

Bedrock Grounding Check

Question 12

What is the primary advantage of the transformer architecture over previous RNN-based architectures for natural language processing tasks?

Accepted Answer

It relies on a self-attention mechanism that allows parallel processing of all tokens. The transformer architecture's primary advantage is its self-attention mechanism, which allows the model to process all tokens in the input sequence in parallel rather than sequentially. This parallelization dramatically reduces training time and enables the model to capture long-range dependencies more effectively than RNNs, which must process tokens one by one and suffer from vanishing gradients over long sequences.

Answer

It processes tokens sequentially to maintain the order of the sequence

Answer

It requires less training data because it uses convolutional layers

Answer

It uses a gated recurrent unit to selectively forget information

Question 13

A company needs to generate high-quality product images from textual descriptions for an e-commerce catalog. They want to use a foundation model on AWS that specializes in text-to-image generation. Which model provider should they use through Amazon Bedrock?

Accepted Answer

Stability AI. Stability AI is the correct provider because its Stable Diffusion models, available through Amazon Bedrock, are specifically designed and optimized for text-to-image generation. These models use a diffusion process to generate high-quality images from textual descriptions, making them ideal for the e-commerce catalog use case.

Answer