AIF-C01 Practice Test 45 — 15 Questions

Question 1

A company uses Amazon Bedrock Agents for customer support. The agent needs to perform multi-step reasoning: first identify the customer's account, then check order status, and finally provide a resolution. Which THREE components must be configured to enable this workflow? (Select THREE.)

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

Accepted Answer

Agent orchestration to plan and execute multi-step reasoning. Option A is correct because Amazon Bedrock Agents rely on the agent's orchestration (the ReAct-based prompt/orchestration layer) to decompose the user request into a plan and sequence the multi-step reasoning across account identification, order-status lookup, and resolution. Option C is correct because each action group is backed by a Lambda function that Bedrock invokes to actually execute the business logic/API calls (for example, calling the CRM or order-management API) and return results to the agent. Option E is correct because you must define an action group per API operation—such as one for account lookup and one for order status—with its OpenAPI schema so the agent knows the available actions, parameters, and when to invoke them. Option B is not required: a Knowledge Base is for retrieval-augmented generation over documents, not for executing the transactional API calls this workflow needs. Option D is not required: a Guardrail adds safety/content filtering but does not enable the multi-step action execution workflow.

Answer

A Bedrock Knowledge Base with customer data

Answer

A Bedrock Guardrail to block out-of-scope questions

Question 2

A binary classification model outputs probabilities. The default threshold of 0.5 results in high precision but low recall. Which action would likely increase recall while maintaining acceptable precision?

Accepted Answer

Decrease the threshold to 0.3. Decreasing the threshold to 0.3 makes the model classify more instances as positive, which increases recall (more true positives captured) but may also increase false positives. The goal is to shift the precision-recall trade-off toward higher recall while keeping precision at an acceptable level, which is directly achieved by lowering the decision threshold.

Answer

Use F1 score instead of accuracy

Answer

Apply oversampling to the minority class

Answer

Increase the threshold to 0.7

Question 3

A data science team is deploying a model using Amazon SageMaker. They need to monitor the model for bias after it is deployed. Which AWS service or feature should they use?

Accepted Answer

Amazon SageMaker Clarify. Amazon SageMaker Clarify provides bias detection and explainability for machine learning models, including pre-training bias metrics on datasets and post-training bias metrics on deployed model predictions. It integrates with SageMaker Model Monitor to continuously detect bias drift in production.

Answer

Amazon Bedrock Guardrails

Answer

Amazon SageMaker Model Monitor

Answer

AWS CloudTrail

Question 4

A startup needs to generate product descriptions from bullet points using a foundation model. They want a fully managed serverless experience. Which AWS service should they use?

Accepted Answer

Amazon Bedrock. Amazon Bedrock is a fully managed serverless service that provides access to foundation models (FMs) from leading AI providers via an API, making it ideal for generating product descriptions from bullet points. It eliminates infrastructure management while allowing you to invoke models like Anthropic Claude or Amazon Titan for text generation tasks.

Answer

Amazon Comprehend

Answer

Amazon Polly

Answer

Amazon Lex

Question 5

Which TWO techniques provide interpretability for machine learning models at a local (per-prediction) level? (Choose two.)

Accepted Answer

SHAP values. SHAP (SHapley Additive exPlanations) values (option A) are correct because they decompose an individual prediction into additive per-feature contributions based on Shapley values from cooperative game theory, giving a local, per-instance explanation of how each feature pushed the model output. LIME (Local Interpretable Model-agnostic Explanations) (option D) is also correct because it fits a simple interpretable surrogate model (e.g., sparse linear model) around a single prediction by perturbing the instance's neighborhood, thereby explaining that specific prediction locally. Partial dependence plots (option B) are not local — they show the average marginal effect of a feature across the entire dataset, a global technique. A confusion matrix (option C) is a global performance summary of classification outcomes (TP/FP/TN/FN counts), not a per-prediction explanation. Permutation feature importance (option E) is a global method that measures the drop in overall model performance when a feature's values are shuffled across the dataset, so it does not explain individual predictions.

Answer

Partial dependence plots

Answer

Confusion matrix

Answer

Permutation feature importance

Question 6

A company is using Amazon Textract to extract text from scanned documents stored in an S3 bucket. The security team requires that all access to the documents be logged and that the documents be encrypted at rest using a customer-managed key. What should the company do to meet these requirements?

Accepted Answer

Enable S3 server-side encryption with AWS KMS (SSE-KMS) and enable CloudTrail data events for the S3 bucket. Enabling S3 server-side encryption with AWS KMS (SSE-KMS) satisfies the requirement for encryption at rest using a customer-managed key, and enabling CloudTrail data events for the S3 bucket captures all access to the documents (including GetObject, PutObject, etc.) for logging. This combination meets both security requirements precisely.

Answer

Use S3 default encryption and enable Textract logging

Answer

Enable S3 server access logs and use S3 SSE-KMS

Answer

Use S3 server-side encryption with S3-managed keys (SSE-S3) and enable S3 access logs

Question 7

Refer to the exhibit. A developer sees this error when calling Amazon Bedrock for inference. What is the MOST likely cause and recommended solution?

Accepted Answer

The request rate exceeds the model's throughput limit; implement retries with exponential backoff. The error indicates a throttling exception from Amazon Bedrock, which occurs when the request rate exceeds the model's throughput limit. The recommended solution is to implement retries with exponential backoff to handle transient rate limits gracefully, as this aligns with AWS best practices for managing API call limits.

Answer

The model ID is incorrect; use a different model

Answer

The prompt is too long; reduce the number of tokens in the prompt

Answer

Increase the max_tokens_to_sample value

Question 8

A team is building an AI chatbot that will be used by customers with visual impairments. Which design practice best supports inclusive accessibility for this user group?

Accepted Answer

Ensure the chatbot is fully navigable via keyboard and compatible with screen readers. Ensuring full keyboard navigability and compatibility with screen readers is a fundamental accessibility requirement for users with visual impairments, enabling them to interact without relying on visual cues.

Answer

Add captions to all video content

Answer

Provide a voice-only interface without any text output

Answer

Use high-contrast color schemes and large fonts

Question 9

An administrator reviews a CloudTrail log entry for a CreateModel API call. Which security concern should they investigate?

Accepted Answer

The model data URL uses HTTP instead of HTTPS. Using HTTP instead of HTTPS for the model data URL exposes the data in transit to potential interception or tampering. CloudTrail logs record the URL as provided, and if it uses HTTP, the data transferred from that URL to SageMaker is unencrypted, violating security best practices for data in transit. This is a direct security concern that should be investigated and remediated by using HTTPS.

Answer

The model name is not encrypted in the log

Answer

The source IP address is an external IP

Answer

The execution role ARN is visible in the log

Question 10

A company is using Amazon Comprehend to extract entities from customer support tickets. The compliance team requires that the text sent to Comprehend be encrypted in transit and that Comprehend does not store any data beyond the processing time. How should the company configure the API call?

Accepted Answer

Use the HTTPS endpoint for the DetectEntities API and rely on Comprehend's stateless design. Amazon Comprehend's DetectEntities API is inherently stateless—it does not store any text or results beyond the processing time required to generate the response. By using the HTTPS endpoint, the company ensures encryption in transit via TLS, satisfying the compliance requirement without additional configuration. Comprehend does not persist data from synchronous API calls, so no data is retained after the response is returned.

Answer

Encrypt the text using AWS KMS before sending it to Comprehend

Answer

Use the AWS SDK with server-side encryption enabled for the API call

Answer

Create a VPC endpoint for Comprehend and send requests over the private network

Question 11

A company wants to automatically detect and redact personally identifiable information (PII) from customer support transcripts. Which TWO AWS services can be used together to achieve this? (Choose two.)

Accepted Answer

Amazon Comprehend. Amazon Transcribe [CORRECT] is the right service to convert the customer support audio transcripts into text, and it also supports PII redaction directly in its transcription jobs via the ContentRedaction parameter with PII redaction type, which detects and masks sensitive data such as names, addresses, and credit card numbers. Amazon Comprehend [CORRECT] is the natural language processing service that provides PII detection and redaction (using the ContainsPiiEntities API and the PII redaction feature) for text, so it can be used together with Transcribe to automatically identify and redact PII in the resulting transcripts. Amazon Rekognition is for image and video analysis (facial recognition, object detection), not text PII redaction, so it does not fit. Amazon Textract extracts text and data from scanned documents (OCR), which is unnecessary here since the source is support transcripts, not document images. Amazon Polly is a text-to-speech service that converts text into lifelike speech, the opposite of what is needed for detecting and redacting PII in transcripts.

Answer

Amazon Rekognition

Answer

Amazon Textract

Answer

Amazon Polly

Question 12

An AI practitioner is fine-tuning an Amazon Titan Text model on a dataset of customer support conversations to improve response accuracy. After training, the model's perplexity on the validation set is low, but during inference, the model frequently generates off-topic or nonsensical responses to real customer queries. What is the most likely cause?

Accepted Answer

The validation set does not represent the distribution of real customer queries. The core issue is a mismatch between the validation set and real-world inference data. Low perplexity on the validation set indicates the model fits that specific distribution well, but if the validation set does not reflect the diversity, phrasing, or intent of actual customer queries, the model will generate off-topic or nonsensical responses during inference. This is a classic case of distribution shift, where the model has not generalized to the true target domain.

Answer

The model's context window is too small for the inference queries

Answer

The temperature during inference is set too low

Answer

The fine-tuning dataset is too small, causing overfitting

Question 13

Which of the following is a key advantage of using a pre-trained foundation model over training a model from scratch?

Accepted Answer

Reduces the amount of labeled data and compute resources needed. Pre-trained foundation models, such as those based on transformer architectures, have already learned general language patterns from vast, diverse datasets during their initial training. This transfer learning approach drastically reduces the need for large amounts of labeled data and extensive compute resources when adapting the model to a specific downstream task, as only a relatively small fine-tuning step is required.

Answer

Eliminates the need for any fine-tuning

Answer

Guarantees perfect accuracy on domain-specific tasks

Answer

Allows the model to work offline without any cloud infrastructure

Question 14

A developer is using the Amazon Bedrock InvokeModel API with a model that has a context window of 8,000 tokens. The developer sends a prompt that is 7,500 tokens long and expects a response of about 1,000 tokens. The API call fails with an error indicating the input exceeds the model's context window. Why did this happen?

Accepted Answer

The context window includes both input and output tokens, so the total of 7,500 + 1,000 = 8,500 exceeds the 8,000 limit. The context window of a foundation model includes both the input prompt and the generated output tokens. The developer's prompt of 7,500 tokens plus the expected 1,000-token response totals 8,500 tokens, which exceeds the model's 8,000-token limit. The InvokeModel API enforces this combined limit, so even though the prompt alone is under the limit, the total token count causes the error.

Answer

The API automatically reserves 2,000 tokens for output, reducing available input capacity

Answer

The model requires a minimum of 512 tokens for internal processing

Answer

The model's context window counts only input tokens; output tokens are separate

Question 15

A company is using Amazon Bedrock with a foundation model for a text summarization task. They want to evaluate the quality of the summaries. Which TWO metrics are appropriate for evaluating the quality of generated summaries? (Select TWO)

Accepted Answer

ROUGE score. ROUGE (Recall-Oriented Understudy for Gisting Evaluation) is correct because it measures n-gram and longest-common-subsequence overlap between generated summaries and reference summaries, which directly captures summary content coverage and is the standard metric for summarization quality. BLEU is also correct because it computes precision-based n-gram overlap between generated and reference text, and while traditionally used for machine translation, it is widely applied to summarization evaluation to assess how much of the generated summary matches the reference phrasing. Accuracy is not appropriate because summarization is a free-form generation task without a single discrete correct label, so classification-style accuracy cannot be computed meaningfully. Latency measures inference speed, not output quality, so it does not evaluate the summary's content. Perplexity measures how well a language model predicts a token sequence and reflects fluency/likelihood, but it does not compare generated summaries against references and thus is not a direct quality metric for summarization.

Answer

Accuracy

Answer

Latency

Answer

Perplexity