AIF-C01 Practice Test 28 — 15 Questions

Question 1

A developer implements a RAG pipeline with Amazon Bedrock and Amazon OpenSearch Serverless. Users report that the chatbot sometimes returns off-topic responses. Investigation shows the retrieved chunks are semantically unrelated to the queries. What is the MOST likely cause?

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

Accepted Answer

The chunking strategy is not optimised for the document type, or the embedding model is not well-suited. Off-topic responses in a RAG pipeline almost always trace back to poor retrieval quality, which is driven by how documents are chunked and which embedding model is used. If chunks are too large, too small, or split mid-sentence, or if the embedding model was trained on a different domain than the corpus, the vector similarity search returns semantically unrelated chunks that the LLM then hallucinates around. Tuning chunk size/overlap and selecting a domain-appropriate embedding model (e.g., Titan Embeddings vs. Cohere Embed) directly fixes retrieval relevance.

Answer

The LLM used for generation has too few parameters

Answer

The agent's action group has misconfigured Lambda functions

Answer

The vector store index is not encrypted at rest

Question 2

A financial services company is deploying a foundation model on Amazon Bedrock to generate compliance reports from internal audit logs. The model must not output any personally identifiable information (PII). They have configured a Bedrock Guardrail with sensitive information filters set to the 'HIGH' sensitivity level. During testing in a staging environment, testers still observed PII being occasionally generated in the report outputs. The guardrail did not block these instances because the PII was embedded in a context that the guardrail's pattern matching did not catch (e.g., structured JSON data with embedded names). The company requires a solution that minimizes latency and cost, as they process thousands of reports daily. They cannot afford to increase inference time significantly due to strict SLAs. They also want to avoid re-engineering the entire solution. Which additional step should they take to effectively eliminate PII leakage while maintaining performance?

Accepted Answer

Implement a post-processing Lambda function that uses Amazon Comprehend's PII detection to scan and redact any PII from the model output before returning it.. Amazon Comprehend's PII detection API can be invoked as a post-processing step to scan and redact PII from the model output without requiring any changes to the model or guardrail configuration. This approach adds minimal latency (typically under 100ms per request) and cost per API call is low, making it suitable for high-throughput scenarios. It directly addresses the guardrail's failure to catch PII embedded in structured contexts like JSON, as Comprehend uses machine learning models that can identify PII even when it's not in plain text patterns.

Answer

Add a prompt instruction to the model to never output PII, with few-shot examples of non-PII outputs.

Answer

Fine-tune the foundation model on a dataset that excludes PII.

Answer

Increase the guardrail sensitivity to 'MAXIMUM'.

Question 3

Which AWS service is best suited for extracting text (including handwriting) from scanned documents, such as invoices and forms?

Accepted Answer

Amazon Textract. Amazon Textract is purpose-built to extract text, handwriting, and structured data (like tables and forms) from scanned documents. It uses machine learning to go beyond simple OCR by understanding the layout and relationships within documents, making it ideal for processing invoices and forms.

Answer

Amazon Comprehend

Answer

Amazon Transcribe

Answer

Amazon Rekognition

Question 4

Which AWS service provides a serverless experience for building and scaling generative AI applications with access to various foundation models?

Accepted Answer

Amazon Bedrock. Amazon Bedrock is a fully managed service that provides a serverless experience for building and scaling generative AI applications. It offers access to a variety of foundation models (FMs) from providers like AI21 Labs, Anthropic, Cohere, Meta, Stability AI, and Amazon via a single API, without the need to manage underlying infrastructure.

Answer

Amazon SageMaker

Answer

Amazon Lex

Answer

AWS Lambda

Question 5

An organization wants to use Amazon Bedrock to generate personalized email content for marketing campaigns. They have a large dataset of customer profiles. Which approach would be MOST cost-effective for high-volume generation?

Accepted Answer

Use batch inference to generate emails asynchronously in large batches. Batch inference in Amazon Bedrock processes large volumes of inference requests asynchronously, which reduces per-request cost compared to real-time invocation and is ideal for high-volume, non-latency-sensitive tasks like generating personalized email content. This approach leverages Bedrock's batch processing capabilities to optimize throughput and cost efficiency.

Answer

Fine-tune the model on past successful emails

Answer

Enable model caching to store previously generated emails

Answer

Use a larger model for better quality

Question 6

A company wants to build a generative AI application that can automatically classify customer feedback into positive, neutral, or negative sentiments. Which foundation model capability is BEST suited for this task?

Accepted Answer

Text classification. Text classification is the foundation model capability specifically designed to categorize text into predefined classes, such as positive, neutral, or negative sentiment. It is the best fit for sentiment analysis because it directly maps input text to a discrete label set.

Answer

Translation

Answer

Text generation

Answer

Text summarization

Question 7

A startup wants to integrate a generative AI chatbot into their mobile app with minimal latency. Which AWS service is purpose-built for deploying foundation models with low latency and high throughput?

Accepted Answer

Amazon Bedrock. Amazon Bedrock is a fully managed service that provides access to foundation models (FMs) from leading AI providers via a serverless API, purpose-built for deploying generative AI applications with low latency and high throughput. It handles the underlying infrastructure, model hosting, and scaling automatically, making it ideal for integrating a generative AI chatbot into a mobile app with minimal latency.

Answer

AWS Lambda

Answer

Amazon SageMaker

Answer

Amazon Transcribe

Question 8

Which Amazon Bedrock model provider offers the Titan family of models, including Titan Text Lite, Titan Text Express, and Titan Image Generator?

Accepted Answer

Amazon. Amazon Bedrock's Titan family of models, including Titan Text Lite, Titan Text Express, and Titan Image Generator, is developed and provided exclusively by Amazon. These models are built on Amazon's own research and infrastructure, designed for tasks such as text generation, summarization, and image creation, and are distinct from third-party providers like Stability AI, Anthropic, or Cohere.

Answer

Stability AI

Answer

Anthropic

Answer

Cohere

Question 9

A social media platform uses an AI system to moderate content. They want to ensure that human reviewers can review decisions when the AI is uncertain. Which AWS service can be used to set up a human review workflow for AI predictions?

Accepted Answer

Amazon Augmented AI (A2I). Amazon Augmented AI (A2I) is purpose-built to integrate human review into ML workflows, allowing you to define review workflows that trigger when model confidence is low. It integrates natively with Amazon Rekognition, Textract, and custom models via the A2I API, and routes uncertain predictions to human reviewers. This directly matches the requirement for human review of AI moderation decisions.

Answer

Amazon Mechanical Turk

Answer

AWS Lambda

Answer

Amazon SageMaker Ground Truth

Question 10

A company is developing an AI system that transcribes medical consultations. To ensure privacy and security, they need to implement controls that protect patient health information (PHI). Which AWS service can help anonymize data before it is used for model training?

Accepted Answer

Amazon Comprehend Medical. Amazon Comprehend Medical is a HIPAA-eligible NLP service purpose-built for extracting and detecting PHI from unstructured medical text such as clinical notes and transcriptions. It includes a DetectPHI API that identifies 18+ PHI categories (names, dates, IDs, contact info) so they can be redacted or anonymized before the data is used for model training.

Answer

Amazon SageMaker Ground Truth

Answer

AWS Lake Formation

Answer

Amazon Rekognition

Question 11

A data engineer needs to ensure that all data uploaded to an S3 bucket for SageMaker training is automatically encrypted with a customer-managed key. Which S3 feature should they enable?

Accepted Answer

Default encryption with SSE-KMS using a customer managed key (CMK).. The requirement specifies a customer-managed key (CMK). Default encryption with SSE-KMS allows you to specify a CMK, ensuring all objects uploaded to the S3 bucket are automatically encrypted using that key. This satisfies the data engineer's need for control over the encryption key.

Answer

Object lock with compliance mode.

Answer

Default encryption with SSE-KMS using an AWS managed key.

Answer

Default encryption with SSE-S3.

Question 12

A healthcare organization is deploying an AI system to assist in diagnosing diseases from medical images. They need to ensure the system is robust, safe, and subject to human oversight. Which TWO actions align with responsible AI guidelines? (Select TWO.)

Accepted Answer

Monitor model performance over time to detect data drift. Option A is correct because continuously monitoring model performance over time to detect data drift is essential for maintaining robustness and safety in a clinical AI system, since changes in patient demographics, imaging equipment, or disease prevalence can degrade accuracy and introduce risk. Option C is correct because Amazon Augmented AI (A2I) enables human review workflows for low-confidence or uncertain predictions, directly satisfying the requirement for meaningful human oversight in high-stakes diagnostic decisions. Options B, D, and E do not belong: deploying directly to production without testing violates safety and validation practices, permanently disabling human review removes the required oversight, and relying on only one source of training data undermines robustness and generalizability across diverse patient populations.

Answer

Deploy the model directly to production without testing

Answer

Permanently disable human review to speed up diagnoses

Answer

Use only one source of training data

Question 13

A data scientist needs to predict whether a transaction is fraudulent (Yes/No). Which type of machine learning problem is this?

Accepted Answer

Classification. This is a classification problem because the output is a discrete category (fraudulent or not). The data scientist is predicting a binary label (Yes/No), which is the defining characteristic of binary classification in supervised learning.

Answer

Clustering

Answer

Regression

Answer

Reinforcement learning

Question 14

A company wants to identify customer segments based on purchasing behavior. They have unlabeled transaction data and do not know the segment definitions beforehand. Which type of machine learning should they use?

Accepted Answer

Unsupervised learning. Unsupervised learning is the correct choice because the company has unlabeled transaction data and no predefined segment definitions. This type of machine learning discovers hidden patterns, groupings, or structures in data without requiring labeled outputs, making it ideal for customer segmentation tasks like clustering.

Answer

Semi-supervised learning

Answer

Supervised learning

Answer

Reinforcement learning

Question 15

A data science team has trained a gradient boosting model using Amazon SageMaker to predict equipment failures. The model's confusion matrix shows 100 true negatives, 5 false positives, 20 false negatives, and 75 true positives. The cost of a false negative (missed failure) is $10,000, and the cost of a false positive (false alarm) is $500. What is the total cost of the model's predictions on this evaluation set?

Accepted Answer

$202,500. The total cost is calculated by multiplying the number of false negatives (20) by their unit cost ($10,000) and the number of false positives (5) by their unit cost ($500), then summing these values: (20 × $10,000) + (5 × $500) = $200,000 + $2,500 = $202,500. This matches option D. The true negatives and true positives have no associated cost in this evaluation.

Answer

$2,500

Answer

$20,500

Answer

$1,020,500