AIP Question Bank · Latest 50 Questions Free Preview (Page 3/5)

Questions 21–30 of the latest 50 AIP questions (5 pages total), with answers and explanations, no login required. Practice all 98 questions online.

Question 21 · No. 68 · Single choice

A financial services company is developing a generative AI (GenAI) application that serves both premium customers and standard customers. The application uses AWS Lambda functions behind an Amazon API Gateway REST API to process requests. The company needs to dynamically switch between AI models based on which customer tier each user belongs to. The company also wants to perform A/B testing for new features without redeploying code. The company needs to validate model parameters like temperature and maximum token limits before applying changes. Which solution will meet these requirements with the LEAST operational overhead?

A. Create an AWS Systems Manager Parameter Store parameters for each configuration. Use Lambda functions to poll for parameter updates. Use Amazon EventBridge events to trigger redeployments when configurations change.
B. Store model configurations in Amazon DynamoDB tables. Optimize access patterns to retrieve configurations according to customer tier. Configure Lambda functions to query DynamoDB at the beginning of each request to determine which model to use.
C. Use AWS AppConfig to manage model configurations. Use feature flags to perform A/B testing. Define JSON schema validation rules for model parameters. Configure Lambda functions to retrieve configurations by using the AWS AppConfig Agent.
D. Create an Amazon ElastiCache (Redis OSS) cluster to store model configurations. Set short TTL values. Run custom validation logic in Lambda functions. Use Amazon CloudWatch metrics to monitor configuration usage.

Answer: C

Explanation: AWS AppConfig is purpose-built for managing feature flags and dynamic configurations, which directly addresses all the requirements. It supports A/B testing through feature flags, provides JSON schema validation for model parameters like temperature and max tokens, and allows Lambda functions to retrieve configurations dynamically via the AppConfig Agent without redeployment. This eliminates polling overhead (Option A), avoids per-request database queries (Option B), and removes the need to manage additional infrastructure like Redis clusters with custom validation logic (Option D). Therefore, Option C delivers the least operational overhead while meeting all stated requirements.

Question 22 · No. 69 · Single choice

A GenAI developer is building a Retrieval Augmented Generation (RAG)-based customer support application that uses Amazon Bedrock foundation models (FMs). The application needs to process 50 GB of historical customer conversations that are stored in an Amazon S3 bucket as JSON files. The application must use the processed data as its retrieval corpus. The application's data processing workflow must extract relevant data from customer support documents, remove customer personally identifiable information (PII), and generate embeddings for vector storage. The processing workflow must be cost-effective and must finish within 4 hours. Which solution will meet these requirements with the LEAST operational overhead?

A. Use AWS Lambda and Amazon Comprehend to process files in parallel, remove PII, and call Amazon Bedrock APIs to generate vectors. Configure Lambda concurrency limits and memory settings to optimize throughput.
B. Create an AWS Glue ETL job to run PII detection scripts on the data. Use Amazon SageMaker Processing to run the HuggingFaceProcessor to generate embeddings by using a pre-trained model. Store the embeddings in Amazon OpenSearch Service.
C. Deploy an Amazon EMR cluster that runs Apache Spark with user-defined functions (UDFs) that call Amazon Comprehend to detect PII. Use Amazon Bedrock APIs to generate vectors. Store outputs in Amazon Aurora PostgreSQL with the pgvector extension.
D. Implement a data processing pipeline that uses AWS Step Functions to orchestrate a workload that uses Amazon Comprehend to detect PII and Amazon Bedrock to generate embeddings. Directly integrate the workflow with Amazon OpenSearch Serverless to store vectors and provide similarity search capabilities.

Answer: D

Explanation: Option D is the best choice because it uses a fully serverless architecture (AWS Step Functions, Amazon Comprehend, Amazon Bedrock, and Amazon OpenSearch Serverless), which requires no infrastructure management. Step Functions orchestrates the workflow, Amazon Comprehend handles PII detection, Amazon Bedrock generates embeddings, and OpenSearch Serverless provides vector storage with built-in similarity search. This approach has the least operational overhead compared to options that require managing Lambda concurrency, Glue/SageMaker coordination, or EMR clusters. It is also cost-effective with pay-per-use pricing and can scale to process 50GB of data within the 4-hour window.

Question 23 · No. 70 · Single choice

A company is using Amazon Bedrock to develop an AI-powered application that uses a foundation model (FM) that supports cross-Region inference and provisioned throughput. The application must serve users in Europe and North America with consistently low latency. The application must comply with data residency regulations that require European user data to remain within Europe-based AWS Regions. During testing, the application experiences service degradation when Regional traffic spikes reach service quotas. The company needs a solution that maintains application resilience and minimizes operational complexity. Which solution will meet these requirements?

A. Deploy separate Amazon Bedrock instances in North American and European Regions. Use a custom routing layer that directs traffic based on user location. Configure Amazon CloudWatch alarms to monitor Regional service usage. Use Amazon SNS to send email alerts to the company when usage approaches specified thresholds.
B. Use Amazon Bedrock cross-Region inference profiles by specifying geographical codes in profile IDs when the application calls the InvokeModel API. Configure separate Amazon API Gateway HTTP APIs to direct European and North American users to the appropriate Regional endpoints.
C. Deploy a multi-Region Amazon API Gateway HTTP API and AWS Lambda functions that implement retry logic to handle throttling. Configure the Lambda functions to call the FM in the nearest secondary Region when the application reaches service quotas in the primary Region. Use intelligent routing to ensure compliance with data residency requirements.
D. Configure provisioned throughput for Amazon Bedrock in multiple Regions. Implement failover logic in the application code to switch between Regions when throttling occurs. Use AWS Global Accelerator to route traffic to the appropriate endpoints based on user location.

Answer: B

Explanation: Option B is the best solution because it leverages Amazon Bedrock's native cross-Region inference profiles feature, which automatically distributes inference requests across multiple Regions to handle traffic spikes without requiring custom code. By specifying geographical codes (e.g., 'us' or 'eu') in the profile IDs, the company ensures that European user data stays within European AWS Regions, complying with data residency regulations. The separate API Gateway HTTP APIs direct users to the appropriate Regional endpoints based on their location, providing low latency. This approach minimizes operational complexity compared to the other options, which require custom routing layers (A), Lambda-based retry logic (C), or custom failover logic with provisioned throughput (D).

Question 24 · No. 71 · Single choice

A financial services company needs to pre-process unstructured data such as customer transcripts, financial reports, and documentation. The company stores the unstructured data in Amazon S3 to support an Amazon Bedrock application. The company must validate data quality, create auditable metadata, monitor data metrics, and customize text chunking to optimize foundation model (FM) performance. Which solution will meet these requirements with the LEAST development effort?

A. Use Amazon SageMaker Data Wrangler to create a data flow. Configure Amazon CloudWatch metrics and alarms to monitor data quality. Use a custom AWS Lambda function to pre-process the data. Load processed data into Amazon Bedrock.
B. Set up an AWS Glue crawler to catalog data sources. Create AWS Glue ETL jobs to run custom transformation scripts. Use AWS Glue Data Quality to validate and monitor data quality. Load processed data into Amazon Bedrock.
C. Use Amazon Comprehend to extract entities. Create an AWS Lambda function to chunk text. Run Amazon Athena to query and validate data quality. Load processed data into Amazon Bedrock.
D. Create an AWS Step Functions workflow to orchestrate data pre-processing tasks. Run custom code on Amazon EC2 instances to process the data. Use Amazon SageMaker Model Monitor to monitor data quality. Load processed data into Amazon Bedrock.

Answer: B

Explanation: Option B is the best choice for meeting the requirements with the least development effort. AWS Glue crawler automatically catalogs data sources and creates auditable metadata. AWS Glue Data Quality is a managed service specifically designed to validate and monitor data quality without requiring extensive custom code. AWS Glue ETL jobs can run custom transformation scripts for text chunking optimization. This integrated solution provides all required capabilities (data quality validation, auditable metadata, data metrics monitoring, and custom text chunking) while minimizing custom development work. The other options either require more development effort (Lambda functions, EC2 custom code) or use services not optimized for the requirements (CloudWatch for data quality, SageMaker Model Monitor for ML models, Athena for quality validation).

Question 25 · No. 72 · Single choice

A company is building a generative AI (GenAI) application that produces content based on a variety of internal and external data sources. The company wants to ensure that the generated output is fully traceable. The application must support data source registration and enable metadata tagging to attribute content to its original source. The application must also maintain audit logs of data access and usage throughout the pipeline. Which solution will meet these requirements?

A. Use AWS Lake Formation to catalog data sources and control access. Apply metadata tags directly in Amazon S3. Use AWS CloudTrail to monitor API activity.
B. Use AWS Glue Data Catalog to register and tag data sources. Use Amazon CloudWatch Logs to monitor access patterns and application behavior.
C. Store data in Amazon S3 and use object tagging for attribution. Use AWS Glue Data Catalog to manage schema information. Use AWS CloudTrail to log access to S3 buckets.
D. Use AWS Glue Data Catalog to register all data sources. Apply metadata tags to attribute data sources. Use AWS CloudTrail to log access and activity across services.

Answer: D

Explanation: Option D is the best solution because it addresses all three requirements comprehensively: (1) AWS Glue Data Catalog is specifically designed for registering and cataloging data sources, (2) Metadata tags applied through the Glue Data Catalog enable proper attribution of generated content to its original source, which is critical for traceability in a GenAI application, and (3) AWS CloudTrail provides comprehensive audit logging across multiple AWS services throughout the entire data pipeline. Option A uses Lake Formation which is more focused on data lake access control rather than content attribution. Option B uses CloudWatch Logs which is better suited for application-level monitoring rather than comprehensive audit trails across services. Option C is fragmented - it separates schema management and tagging into different services and only logs S3 access rather than the entire pipeline activity.

Question 26 · No. 73 · Single choice

An ecommerce company is developing a generative AI (GenAI) solution that uses Amazon Bedrock with Anthropic Claude to recommend products to customers. Customers report that some of the recommended products are not available for sale on the website or are not relevant to the customer. Customers also report that the solutions takes a long time to generate some recommendations. The company investigates the issues and finds that most interactions between customers and the product recommendation solution are unique. The company confirms that the solutions recommends products that are not in the company's product catalog. The company must resolve these issues. Which solution will meet this requirement?

A. Increase grounding within Amazon Bedrock Guardrails. Enable Automated Reasoning checks. Set up provisioned throughput.
B. Use prompt engineering to restrict the model responses to relevant products. Use streaming techniques such as the InvokeModelWithResponseStream action to reduce perceived latency for the customers.
C. Create an Amazon Bedrock knowledge base. Implement Retrieval Augmented Generation (RAG). Set the PerformanceConfigLatency parameter to optimized.
D. Store product catalog data in Amazon OpenSearch Service. Validate the model's product recommendations against the product catalog. Use Amazon DynamoDB to implement response caching.

Answer: C

Explanation: The best solution is C because it addresses all three issues: (1) Creating an Amazon Bedrock knowledge base and implementing Retrieval Augmented Generation (RAG) ensures the model retrieves actual product information from the company's catalog, directly solving the problem of recommending products not available for sale or not relevant. (2) Setting the PerformanceConfigLatency parameter to optimized helps reduce the latency issue. Option A is incorrect because Automated Reasoning checks are designed for logical verification, not product catalog validation. Option B is insufficient because prompt engineering alone cannot reliably restrict the model to only existing products, and streaming only reduces perceived latency, not actual generation time. Option D is less ideal because caching with DynamoDB won't help when most interactions are unique (as stated in the problem), and validating after generation is less efficient than using RAG to ground the responses from the start.

Question 27 · No. 74 · Single choice

A healthcare company is using Amazon Bedrock to develop a real-time patient care AI assistant to respond to queries for separate departments that handle clinical inquiries, insurance verification, appointment scheduling, and insurance claims. The company wants to use a multi-agent architecture. The company must ensure that the AI assistant is scalable and can onboard new features for patients. The AI assistant must be able to handle thousands of parallel patient interactions. The company must ensure that patients receive appropriate domain-specific responses to queries. Which solution will meet these requirements?

A. Isolate data for each agent by using separate knowledge bases. Use IAM filtering to control access to each knowledge base. Deploy a supervisor agent to perform natural language intent classification on patient inquiries. Configure the supervisor agent to route queries to specialized collaborator agents to respond to department-specific queries. Configure each specialized collaborator agent to use Retrieval Augmented Generation (RAG) with the agent's department-specific knowledge base.
B. Create a separate supervisor agent for each department. Configure individual collaborator agents to perform natural language intent classification for each specialty domain within each department. Integrate each collaborator agent with department-specific knowledge bases only. Implement manual handoff processes between the supervisor agents.
C. Isolate data for each department in separate knowledge bases. Use IAM filtering to control access to each knowledge base. Deploy a single general-purpose agent. Configure multiple action groups within the general-purpose agent to perform specific department functions. Implement rule-based routing logic within the general-purpose agent instructions.
D. Implement multiple independent supervisor agents that run in parallel to respond to patient inquiries for each department. Configure multiple collaborator agents for each supervisor agent. Integrate all agents with the same knowledge base. Use external routing logic to merge responses from multiple supervisor agents.

Answer: A

Explanation: Option A best meets all the requirements. It uses a multi-agent architecture with a supervisor agent that performs intent classification and routes queries to specialized collaborator agents. Each collaborator agent uses RAG with its own department-specific knowledge base, ensuring domain-specific responses. The data isolation through separate knowledge bases with IAM filtering provides security and scalability. This modular design allows easy onboarding of new features/departments by adding new collaborator agents without disrupting existing ones, and the distributed agent architecture can handle thousands of parallel interactions efficiently.

Question 28 · No. 75 · Single choice

A financial technology company is using Amazon Bedrock to build an assessment system for the company's customer service AI assistant. The AI assistant must provide financial recommendations that are factually accurate, compliant with financial regulations, and conversationally appropriate. The company needs to combine automated quality evaluations at scale with targeted human reviews of critical interactions. What solution will meet these requirements?

A. Configure a pipeline in which financial experts manually score all responses for accuracy, compliance, and conversational quality. Use Amazon SageMaker notebooks to analyze results to identify improvement areas.
B. Configure Amazon Bedrock evaluations that use Anthropic Claude Sonnet as a judge model to assess response accuracy and appropriateness. Configure custom Amazon Bedrock guardrails to check responses for compliance with financial policies. Add Amazon Augmented AI (Amazon A2I) human reviews for flagged critical interactions.
C. Create an Amazon Lex bot to manage the customer service interactions. Configure AWS Lambda functions to check responses against a static compliance database. Configure intents in the bot that call the Lambda functions to check the responses. Add an additional intent to collect end-user reviews.
D. Configure Amazon CloudWatch to monitor response patterns from the AI assistant. Configure CloudWatch alerts for potential compliance violations. Establish a team of human evaluators to review flagged interactions.

Answer: B

Explanation: Option B is the correct answer because it directly addresses all the stated requirements. The company needs to combine automated quality evaluations at scale with targeted human reviews of critical interactions while using Amazon Bedrock. Option B uses Amazon Bedrock evaluations with Anthropic Claude Sonnet as a judge model to perform automated quality evaluations at scale for accuracy and appropriateness. It uses custom Amazon Bedrock guardrails to check compliance with financial policies (financial regulations). And it uses Amazon A2I for targeted human reviews of critical interactions that are flagged. This combination perfectly matches the requirements. Option A fails because manual scoring of all responses does not scale. Option C uses Amazon Lex instead of Amazon Bedrock evaluations and a static compliance database which is limited. Option D uses CloudWatch for monitoring rather than proactive AI-based evaluation and lacks automated quality evaluation at scale.

Question 29 · No. 76 · Multiple choice

An ecommerce company is building an internal platform to develop generative AI applications by using Amazon Bedrock foundation models (FMs). Developers need to select models based on evaluations that are aligned to ecommerce use cases. The platform must display accuracy metrics for text generation and summarization in dashboards. The company has custom ecommerce datasets to use as standardized evaluation inputs. Which combination of steps will meet these requirements with the LEAST operational overhead? (Choose two.)

A. Import the datasets to an Amazon S3 bucket. Provide appropriate IAM permissions and cross-origin resource sharing (CORS) permissions to give the evaluation jobs access to the datasets.
B. Import the datasets to an Amazon S3 bucket. Provide appropriate IAM permissions and a VPC endpoint configuration to give the evaluation jobs access to the datasets.
C. Configure an AWS Lambda function to create model evaluation jobs on a schedule in the Amazon Bedrock console. Provide the URI of the S3 bucket that contains the datasets as an input. Configure the evaluation jobs to measure the real world knowledge (RWK) score for text generation and BERT Score for summarization. Configure a second Lambda function to check the status of the jobs and publish custom logs to Amazon CloudWatch. Create a custom Amazon CloudWatch Logs Insights dashboard.
D. Use Amazon SageMaker Clarify on a schedule to create model evaluation jobs. Use open source frameworks to create and run standardized evaluations. Publish results to Amazon CloudWatch namespaces. Use the word error rate score for text generation and toxicity for summarization as metrics for accuracy. Configure an AWS Lambda function to check the status of the jobs and publish custom logs to CloudWatch. Create a custom Amazon CloudWatch Logs Insights dashboard.
E. Run an Amazon SageMaker AI notebook job on a schedule by using the fmevals or ragas framework to run evaluations that use the datasets in the S3 bucket. Write Python code in the notebook that makes direct InvokeModel API calls to the FMs and processes their responses for evaluation. Publish job status and results to Amazon CloudWatch Logs to measure the real world knowledge (RWK) score for text generation and toxicity for summarization as metrics for accuracy. Create a custom CloudWatch Logs Insights dashboard.

Answer: A, C

Explanation: Option A uses Amazon S3 with IAM and CORS permissions, which is the simplest and least operationally complex way to provide Bedrock evaluation jobs access to the custom ecommerce datasets. CORS is required for Bedrock to access the S3 datasets during evaluation. Option C leverages Amazon Bedrock's native model evaluation jobs feature with appropriate accuracy metrics: Real World Knowledge (RWK) score for text generation and BERT Score for summarization—both are suitable metrics for measuring accuracy in these use cases. Using Lambda functions to schedule jobs and check status, combined with CloudWatch Logs Insights dashboards, minimizes operational overhead by using managed services and built-in integrations. Option B adds unnecessary VPC endpoint configuration complexity. Options D and E use inappropriate metrics (word error rate for text generation and toxicity for summarization are not accuracy metrics) and require more operational overhead through custom code, SageMaker jobs, or notebook management.

Question 30 · No. 77 · Single choice

A healthcare company is using Amazon Bedrock to build a Retrieval Augmented Generation (RAG) application that helps practitioners make clinical decisions. The application must achieve high accuracy for patient information retrievals, identify hallucinations in generated content, and reduce human review costs. Which solution will meet these requirements?

A. Use Amazon Comprehend to analyze and classify RAG responses and to extract medical entities and relationships. Use AWS Step Functions to orchestrate automated evaluations. Configure Amazon CloudWatch metrics to track entity recognition confidence scores. Configure CloudWatch to send an alert when accuracy falls below specified thresholds.
B. Implement automated large language model (LLM)-based evaluations that use a specialized model that is fine-tuned for medical content to assess all responses. Deploy AWS Lambda functions to parallelize evaluations. Publish results to Amazon CloudWatch metrics that track relevance and factual accuracy.
C. Configure Amazon CloudWatch Synthetics to generate test queries that have known answers on a regular schedule, and track model success rates. Set up dashboards that compare synthetic test results against expected outcomes.
D. Deploy a hybrid evaluation system that uses an automated LLM-as-a-judge evaluation to initially screen responses and targeted human reviews for edge cases. Use Amazon SageMaker Feature Store to maintain evaluation datasets. Use a built-in Amazon Bedrock evaluation to track retrieval precision and hallucination rates.

Answer: D

Explanation: Option D is the best solution because it addresses all three requirements comprehensively: (1) High accuracy for patient information retrievals is achieved through the combination of automated LLM-as-a-judge evaluation and targeted human reviews for edge cases, ensuring both broad coverage and quality assurance. (2) Identifying hallucinations is directly handled by using the built-in Amazon Bedrock evaluation to track hallucination rates. (3) Reducing human review costs is achieved through the hybrid approach where automated evaluation initially screens responses, and human reviewers only need to focus on edge cases, significantly reducing the workload. Additionally, SageMaker Feature Store helps maintain and manage evaluation datasets efficiently.

Prev 1 234 5 Next