Read Online AIP-C01 Test Practice Test Questions Exam Dumps [Q25-Q44]

Topic	Details
Topic 1	Testing, Validation, and Troubleshooting: This domain covers evaluating foundation model outputs, implementing quality assurance processes, and troubleshooting GenAI-specific issues including prompts, integrations, and retrieval systems.
Topic 2	Operational Efficiency and Optimization for GenAI Applications: This domain encompasses cost optimization strategies, performance tuning for latency and throughput, and implementing comprehensive monitoring systems for GenAI applications.
Topic 3	Foundation Model Integration, Data Management, and Compliance: This domain covers designing GenAI architectures, selecting and configuring foundation models, building data pipelines and vector stores, implementing retrieval mechanisms, and establishing prompt engineering governance.
Topic 4	Implementation and Integration: This domain focuses on building agentic AI systems, deploying foundation models, integrating GenAI with enterprise systems, implementing FM APIs, and developing applications using AWS tools.
Topic 5	AI Safety, Security, and Governance: This domain addresses input output safety controls, data security and privacy protections, compliance mechanisms, and responsible AI principles including transparency and fairness.

QUESTION 25
A company uses Amazon Bedrock to implement a Retrieval Augmented Generation (RAG)-based system to serve medical information to users. The company needs to compare multiple chunking strategies, evaluate the generation quality of two foundation models (FMs), and enforce quality thresholds for deployment.
Which Amazon Bedrock evaluation configuration will meet these requirements?

Create a retrieve-only evaluation job that uses a supported version of Anthropic Claude Sonnet as the evaluator model. Configure metrics for context relevance and context coverage. Define deployment thresholds in a separate CI/CD pipeline.

Create a retrieve-and-generate evaluation job that uses custom precision-at-k metrics and an LLM-as-a- judge metric with a scale of 1-5. Include each chunking strategy in the evaluation dataset. Use a supported version of Anthropic Claude Sonnet to evaluate responses from both FMs.

Create a separate evaluation job for each chunking strategy and FM combination. Use Amazon Bedrock built-in metrics for correctness and completeness. Manually review scores before deployment approval.

Set up a pipeline that uses multiple retrieve-only evaluation jobs to assess retrieval quality. Create separate evaluation jobs for both FMs that use Amazon Nova Pro as the LLM-as-a-judge model.Evaluate based on faithfulness and citation precision metrics.

QUESTION 26
A financial services company uses multiple foundation models (FMs) through Amazon Bedrock for its generative AI (GenAI) applications. To comply with a new regulation for GenAI use with sensitive financial data, the company needs a token management solution.
The token management solution must proactively alert when applications approach model-specific token limits. The solution must also process more than 5,000 requests each minute and maintain token usage metrics to allocate costs across business units.
Which solution will meet these requirements?

Develop model-specific tokenizers in an AWS Lambda function. Configure the Lambda function to estimate token usage before sending requests to Amazon Bedrock. Configure the Lambda function to publish metrics to Amazon CloudWatch and trigger alarms when requests approach thresholds. Store detailed token usage in Amazon DynamoDB to report costs.

Implement Amazon Bedrock Guardrails with token quota policies. Capture metrics on rejected requests.
Configure Amazon EventBridge rules to trigger notifications based on Amazon Bedrock Guardrails metrics. Use Amazon CloudWatch dashboards to visualize token usage trends across models.

Deploy an Amazon SQS dead-letter queue for failed requests. Configure an AWS Lambda function to analyze token-related failures. Use Amazon CloudWatch Logs Insights to generate reports on token usage patterns based on error logs from Amazon Bedrock API responses.

Use Amazon API Gateway to create a proxy for all Amazon Bedrock API calls. Configure request throttling based on custom usage plans with predefined token quotas. Configure API Gateway to reject requests that will exceed token limits.

QUESTION 27
A research company is developing a GenAI system to produce summaries of technical documents. The company must catalog all data sources in a central location. The company needs a solution that can automatically discover and update data sources. The solution must tag each generated summary with citations as metadata that users can query. The solution must retain tamper-evident, immutable audit logs for every model invocation and store input/output records.
Which solution will meet these requirements?

Use Amazon Comprehend to identify data sources in the documents. Store generated summaries in Amazon S3 and enable S3 Object Lock. Use Amazon CloudWatch metrics to generate reports about application throughput. Do not include logs for each invocation.

Use AWS Glue Data Catalog with crawlers to maintain data sources. Store generated summaries in Amazon S3. Write object tags that include a source ID. Store Amazon Bedrock model invocation logs in Amazon S3. Enable S3 Object Lock on the S3 bucket that stores invocation logs. Use AWS CloudTrail log file integrity validation to provide tamper-evident immutability.

Store application outputs in Amazon DynamoDB. Apply item-level tags that include source attribution.
Write application events to Amazon CloudWatch Logs. Use IAM roles to provide audit traceability.

Use AWS AppConfig feature flags to implement data versioning. Restrict access to the model by using IAM condition keys. Maintain a versioned mapping file of source-to-output relationships in Amazon S3.

Option B is correct because it is the only option that satisfies all requirements: centralized cataloging, automatic data-source discovery, citation metadata, model input/output logging, and tamper-evident immutable audit retention. AWS Glue Data Catalog is designed as a central metadata repository, and AWS Glue crawlers can automatically discover and catalog new or updated data sources. AWS documentation states that crawlers can infer metadata and keep the Data Catalog up to date across sources such as Amazon S3, Amazon RDS, Amazon Redshift, and other supported stores. This directly addresses the requirement to catalog all data sources centrally and update them automatically.
Storing generated summaries in Amazon S3 and writing object tags that include source IDs is appropriate because S3 object tags are queryable metadata that can be used for organization, lifecycle policies, access controls, cost allocation, and downstream indexing. For GenAI traceability, the source ID tag can represent the citation or provenance link from the generated summary back to the original technical document.
Amazon Bedrock model invocation logging is also required here because the company must retain input
/output records for every model invocation. AWS documentation states that model invocation logging can collect invocation logs, model input data, and model output data for Amazon Bedrock invocations, with Amazon S3 as a supported delivery destination.
For immutable retention, S3 Object Lock provides write-once-read-many protection and helps prevent objects from being deleted or overwritten for a fixed period or indefinitely. CloudTrail log file integrity validation adds tamper evidence by using hashing and digital signing to detect whether log files were changed, deleted, or forged after delivery.
Option A explicitly omits logs for each invocation, so it fails the audit requirement. Option C lacks automatic source discovery and immutable audit retention. Option D uses configuration tooling, not a cataloging and audit architecture. Therefore, option B is the complete AWS-aligned solution.

QUESTION 28
A company is designing a solution that uses foundation models (FMs) to support multiple AI workloads.
Some FMs must be invoked on demand and in real time. Other FMs require consistent high-throughput access for batch processing.
The solution must support hybrid deployment patterns and run workloads across cloud infrastructure and on- premises infrastructure to comply with data residency and compliance requirements.
Which combination of steps will meet these requirements? (Select TWO.)

Use AWS Lambda to orchestrate low-latency FM inference by invoking FMs hosted on Amazon SageMaker AI asynchronous endpoints.

Configure provisioned throughput in Amazon Bedrock to ensure consistent performance for high- volume workloads.

Deploy FMs to Amazon SageMaker AI endpoints with support for edge deployment by using Amazon SageMaker Neo. Orchestrate the FMs by using AWS Lambda to support hybrid deployment.

Use Amazon Bedrock with auto-scaling to handle unpredictable traffic surges.

Use Amazon SageMaker JumpStart to host and invoke the FMs.

QUESTION 29
A company has a generative AI (GenAI) application that uses Amazon Bedrock to provide real-time responses to customer queries. The company has noticed intermittent failures with API calls to foundation models (FMs) during peak traffic periods.
The company needs a solution to handle transient errors and provide detailed observability into FM performance. The solution must prevent cascading failures during throttling events and provide distributed tracing across service boundaries to identify latency contributors. The solution must also enable correlation of performance issues with specific FM characteristics.
Which solution will meet these requirements?

Implement a custom retry mechanism with a fixed delay of 1 second between retries. Configure Amazon CloudWatch alarms to monitor the application’s error rates and latency metrics.

Configure the AWS SDK with standard retry mode and exponential backoff with jitter. Use AWS X- Ray tracing with annotations to identify and filter service components.

Implement client-side caching of all FM responses. Add custom logging statements in the application code to record API call durations.

Configure the AWS SDK with adaptive retry mode. Use AWS CloudTrail distributed tracing to monitor throttling events.

Option B best meets the combined resiliency and observability requirements because it applies AWS- recommended retry behavior for transient throttling and enables true distributed tracing across service boundaries. During peak traffic, intermittent failures are commonly caused by throttling and other transient conditions. The AWS SDK standard retry mode provides exponential backoff with jitter, which reduces synchronized retry storms, prevents cascading failures, and improves overall system stability. Jitter is important because it spreads retry attempts over time, reducing load amplification during throttling events.
For observability, AWS X-Ray provides distributed tracing that follows a request across components such as API Gateway or load balancers, application services, and downstream calls to Amazon Bedrock. X-Ray can identify where latency is being introduced and which downstream call is contributing most to end-to-end response time. This is required to “identify latency contributors” and isolate performance issues under load.
The requirement also states that the company must correlate performance issues with specific FM characteristics. X-Ray annotations are designed for this purpose: the application can annotate traces with the model ID, inference parameters, region, or inference profile used. This enables filtering and analysis (for example, comparing latency or error patterns by model, parameter set, or endpoint configuration) without building a separate telemetry system.
Option A’s fixed-delay retries increase synchronized retry behavior and do not provide distributed tracing.
Option C does not prevent cascading failures and cannot provide cross-service tracing. Option D is incorrect because CloudTrail is an audit logging service and does not provide distributed tracing for request latency analysis.
Therefore, Option B provides the correct combination of resilient retries and deep, model-correlated distributed observability for Amazon Bedrock workloads.

QUESTION 30
A retail company runs an application that makes product recommendations to customers on the company’s website. The application uses Amazon Bedrock to generate recommendations by dynamically constructing prompts and sending them to foundation models (FMs). A GenAI developer has deployed an update to the application that instructs the FM to include a specific promotional message when the FM generates a response to prompts. When the developer tests the application, the promotional message does not always appear in the responses. When the promotional message does appear in the responses, it does not always flow with the rest of the text. The GenAI developer must ensure that the promotional message always appears in the FM responses. Which solution will meet this requirement?

Use an Amazon Bedrock Guardrails filter on the prompt. Set the input filter strength to HIGH.

Generate multiple response variants that include the promotional message in different ways. Use a reranker model to select the most coherent version based on relevance to the original prompt.

Run the prompt through Amazon Bedrock. Process the response through Amazon Bedrock AgentCore to add the promotional message. Rerank the results by using the original prompt and the desired message as context.

Reinforce the requirement to include the new promotional message within product recommendations by using an output indicator in prompts to the FM.

QUESTION 31
A publishing company is developing a chat assistant that uses a containerized large language model (LLM) that runs on Amazon SageMaker AI. The architecture consists of an Amazon API Gateway REST API that routes user requests to an AWS Lambda function. The Lambda function invokes a SageMaker AI real-time endpoint that hosts the LLM.
Users report uneven response times. Analytics show that a high number of chats are abandoned after 2 seconds of waiting for the first token. The company wants a solution to ensure that p95 latency is under 800 ms for interactive requests to the chat assistant.
Which combination of solutions will meet this requirement? (Select TWO.)

Enable model preload upon container startup. Implement dynamic batching to process multiple user requests together in a single inference pass.

Select a larger GPU instance type for the SageMaker AI endpoint. Set the minimum number of instances to 0. Continue to perform per-request processing. Lazily load model weights on the first request.

Switch to a multi-model endpoint. Use lazy loading without request batching.

Set the minimum number of instances to greater than 0. Enable response streaming.

Switch to Amazon SageMaker Asynchronous Inference for all requests. Store requests in an Amazon S3 bucket. Set the minimum number of instances to 0.

QUESTION 32
A company is planning to deploy multiple generative AI (GenAI) applications to five independent business units that operate in multiple countries in Europe and the Americas. Each application uses Amazon Bedrock Retrieval Augmented Generation (RAG) patterns with business unit-specific knowledge bases that store terabytes of unstructured data.
The company must establish well-architected, standardized components for security controls, observability practices, and deployment patterns across all the GenAI applications. The components must be reusable, versioned, and governed consistently.
Which solution will meet these requirements?

Configure Amazon API Gateway REST API endpoints for the GenAI applications. Deploy common security, observability, and RAG patterns based on the AWS Well-Architected Generative AI Lens in standardized AWS CloudFormation templates. Use CloudFormation Guard after deployment to validate policy compliance in each business unit.

Create standardized AWS CloudFormation templates to implement security, observability, and RAG patterns based on the AWS Well-Architected Generative AI Lens. Establish a centralized repository for version control. Integrate a CI/CD pipeline with CloudFormation Guard to enforce consistent and repeatable deployments across business units.

Use AWS Service Catalog to define standardized portfolios and versioned products for each business unit. Use the portfolios to enforce security, observability, and RAG patterns based on the AWS Well- Architected Generative AI Lens. Require business units to use the Service Catalog console to deploy resources.

Document security controls, observability requirements, and RAG patterns based on the AWS Well- Architected Generative AI Lens in a shared design document. Use Amazon Macie to enforce deployment. Delegate implementation responsibility to each business unit.

Option B best meets the requirement for reusable, versioned, and consistently governed components across multiple business units because it implements “platform-level standardization” through infrastructure as code plus automated compliance enforcement before deployment. Standardized CloudFormation templates provide reusable building blocks for security controls (identity, networking boundaries, encryption), observability practices (metrics, logs, traces), and RAG deployment patterns (knowledge base integration, ingestion pipelines, retrieval controls). This aligns with AWS guidance to operationalize well-architected patterns through repeatable templates rather than ad hoc implementations.
A centralized repository enables version control, change review, and governance of templates across all five business units. This satisfies the “versioned” and “reusable” requirements and provides a single source of truth for approved architectures. Integrating a CI/CD pipeline ensures that deployments are consistent and automated, reducing drift between business units and Regions.
CloudFormation Guard is most effective when used as a preventive control in the pipeline, not only after deployment. By running Guard rules during build or pre-deploy stages, the organization can enforce mandatory security and observability configurations and block noncompliant changes before they reach production. This supports consistent governance while still enabling business units to deploy quickly.
Option A performs compliance validation after deployment, which allows policy violations to be deployed first and remediated later. Option C provides governed provisioning but requiring console-based deployment reduces automation and can slow standardized CI/CD adoption; it also adds an additional governance layer that is not required to meet the stated needs. Option D is not enforceable and does not provide reusable, versioned, governed components.
Therefore, Option B provides the strongest, most scalable, and most consistently governed approach for standardized GenAI deployments across business units.

QUESTION 33
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?

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.

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 .

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.

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.

QUESTION 34
A media company is building an AI-powered content moderation system by using Amazon Bedrock. The system first classifies text by using a small, low-latency model. Then the system escalates requests that have a confidence score below 0.65 to a larger, more expensive model.
The system must respond in near real time for high-confidence results. The system must process low- confidence requests asynchronously. The system must scale to meet sudden spikes in demand. The company wants to optimize costs for the system by invoking the larger model only when required. The company wants to use decoupled components to achieve high resiliency for the system.
Which solution will meet these requirements?

Use Amazon API Gateway to invoke the small model synchronously. If the small model’s confidence score is below 0.65, synchronously call the larger model. Use provisioned concurrency to handle traffic spikes.

Use an AWS Step Functions workflow that has parallel branches to run both the small model and the large model for every request. Choose the large model result when confidence score values differ.

Send requests to an Amazon SQS queue. Use AWS Fargate to process messages. Invoke the small model first. If the confidence score is below 0.65, place the request in a second SQS queue to process asynchronously by using the large model.

Deploy both models on Amazon EC2 instances and enable auto scaling. Use a custom application heuristic to route requests to the appropriate instance based on phrase length and keyword rules.

Option C is the best answer because it implements a decoupled, queue-based moderation pipeline that invokes the expensive model only when the low-latency model is not confident enough. Amazon SQS is designed to decouple distributed application components and support asynchronous processing. AWS documentation describes SQS as a fully managed message queuing service that enables decoupling and scaling of microservices, distributed systems, and serverless applications. This matches the requirement for high resiliency and sudden demand spikes because incoming requests can be buffered in a durable queue rather than overwhelming the model-processing layer.
Using AWS Fargate to process queue messages is also appropriate because Fargate provides serverless container compute for Amazon ECS or Amazon EKS workloads. It allows the company to run scalable processing workers without managing EC2 capacity directly. AWS Prescriptive Guidance includes architectures that use API Gateway, Amazon SQS, and AWS Fargate to process events asynchronously, which supports the same decoupled processing model required in this question.
The two-stage queue design also optimizes cost. The small, low-latency classifier is used first for all requests.
Only requests with confidence below 0.65 are placed into the second queue and processed by the larger model. This avoids running the larger model for every moderation request. High-confidence results can be completed quickly by the first-stage processor, while uncertain results are isolated into an asynchronous second-stage workflow.
Option A is incorrect because it synchronously calls the larger model for low-confidence results, which violates the requirement to process low-confidence requests asynchronously. Option B is incorrect because it invokes both models for every request, increasing cost and eliminating the benefit of confidence-based escalation. Option D requires managing EC2 instances and uses keyword heuristics instead of model confidence, so it is less resilient and less aligned with Bedrock-based moderation. Therefore, option C is correct.

QUESTION 35
An ecommerce company is developing a generative AI application that uses Amazon Bedrock with Anthropic Claude to recommend products to customers. Customers report that some recommended products are not available for sale on the website or are not relevant to the customer. Customers also report that the solution 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 solution recommends products that are not in the company’s product catalog. The company must resolve these issues.
Which solution will meet this requirement?

Increase grounding within Amazon Bedrock Guardrails. Enable Automated Reasoning checks. Set up provisioned throughput.

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.

Create an Amazon Bedrock knowledge base. Implement Retrieval Augmented Generation RAG. Set the PerformanceConfigLatency parameter to optimized.

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.

Option C best addresses both core problems: hallucinated recommendations that do not exist in the catalog and slow response times, while keeping operational overhead low. The most direct way to prevent the model from recommending unavailable products is to ground generation on authoritative product catalog data at inference time. An Amazon Bedrock knowledge base is designed for this pattern by ingesting domain data, chunking content, creating embeddings, and retrieving the most relevant catalog entries when a user asks for recommendations. Implementing Retrieval Augmented Generation ensures the foundation model receives only approved, catalog-backed context and can cite or base its output on those retrieved items. This sharply reduces the likelihood of inventing products, because the response is conditioned on retrieved catalog records rather than relying on the model’s parametric memory.
The requirement also notes that most interactions are unique. That makes response caching far less effective, because there are fewer repeated prompts to benefit from cached outputs. Instead, improving the retrieval and model invocation path is the better optimization. Using the PerformanceConfigLatency parameter set to optimized prioritizes lower latency behavior for model inference, helping meet faster recommendation generation without requiring the company to build and operate additional infrastructure.
The other options do not solve the root cause as reliably. Prompt engineering and streaming can improve perceived latency, but they do not guarantee catalog-only recommendations because the model can still hallucinate items. Guardrails can help detect or block certain undesired outputs, but without consistent catalog grounding they do not ensure every recommendation is derived from the company’s product data. Building a custom OpenSearch validation and caching layer increases operational complexity, and caching is misaligned with predominantly unique interactions.
Alright, after comparing List B (txt file) against List A (Word file) , I have identified the unique questions.
These questions cover scenarios or architectural configurations that were not present in the existing list.
Here are the unique questions from List B, formatted as requested:

QUESTION 36
A company uses Amazon Bedrock to build a Retrieval Augmented Generation (RAG) system. The RAG system uses an Amazon Bedrock Knowledge Bases that is based on an Amazon S3 bucket as the data source for emergency news video content. The system retrieves transcripts, archived reports, and related documents from the S3 bucket.
The RAG system uses state-of-the-art embedding models and a high-performing retrieval setup. However, users report slow responses and irrelevant results, which cause decreased user satisfaction. The company notices that vector searches are evaluating too many documents across too many content types and over long periods of time.
The company determines that the underlying models will not benefit from additional fine-tuning. The company must improve retrieval accuracy by applying smarter constraints and wants a solution that requires minimal changes to the existing architecture.
Which solution will meet these requirements?

Enhance embeddings by using a domain-adapted model that is specifically trained on emergency news content for improved vector similarity.

Migrate to Amazon OpenSearch Service. Use vector fields and metadata filters to define the scope of results retrieval.

Enable metadata-aware filtering within the Amazon Bedrock knowledge base by indexing S3 object metadata.

Migrate to an Amazon Q Business index to perform structured metadata filtering and document categorization during retrieval.

QUESTION 37
A company uses AWS Lake Formation to set up a data lake that contains databases and tables for multiple business units across multiple AWS Regions. The company wants to use a foundation model (FM) through Amazon Bedrock to perform fraud detection. The FM must ingest sensitive financial data from the data lake.
The data includes some customer personally identifiable information (PII).
The company must design an access control solution that prevents PII from appearing in a production environment. The FM must access only authorized data subsets that have PII redacted from specific data columns. The company must capture audit trails for all data access.
Which solution will meet these requirements?

Create a separate dataset in a separate Amazon S3 bucket for each business unit and Region combination. Configure S3 bucket policies to control access based on IAM roles that are assigned to FM training instances. Use S3 access logs to track data access.

Configure the FM to authenticate by using AWS Identity and Access Management roles and Lake Formation permissions based on LF-Tag expressions. Define business units and Regions as LF-Tags that are assigned to databases and tables. Use AWS CloudTrail to collect comprehensive audit trails of data access.

Use direct IAM principal grants on specific databases and tables in Lake Formation. Create a custom application layer that logs access requests and further filters sensitive columns before sending data to the FM.

Configure the FM to request temporary credentials from AWS Security Token Service. Access the data by using presigned S3 URLs that are generated by an API that applies business unit and Regional filters. Use AWS CloudTrail to collect comprehensive audit trails of data access.

QUESTION 38
A company is implementing a serverless inference API by using AWS Lambda. The API will dynamically invoke multiple AI models hosted on Amazon Bedrock. The company needs to design a solution that can switch between model providers without modifying or redeploying Lambda code in real time. The design must include safe rollout of configuration changes and validation and rollback capabilities.
Which solution will meet these requirements?

Store the active model provider in AWS Systems Manager Parameter Store. Configure a Lambda function to read the parameter at runtime to determine which model to invoke.

Store the active model provider in AWS AppConfig. Configure a Lambda function to read the configuration at runtime to determine which model to invoke.

Configure an Amazon API Gateway REST API to route requests to separate Lambda functions.
Hardcode each Lambda function to a specific model provider. Switch the integration target manually.

Store the active model provider in a JSON file hosted on Amazon S3. Use AWS AppConfig to reference the S3 file as a hosted configuration source. Configure a Lambda function to read the file through AppConfig at runtime to determine which model to invoke.

QUESTION 39
A financial services company is creating a Retrieval Augmented Generation (RAG) application that uses Amazon Bedrock to generate summaries of market activities. The application relies on a vector database that stores a small proprietary dataset with a low index count. The application must perform similarity searches.
The Amazon Bedrock model’s responses must maximize accuracy and maintain high performance.
The company needs to configure the vector database and integrate it with the application.
Which solution will meet these requirements?

Launch an Amazon MemoryDB cluster and configure the index by using the Flat algorithm. Configure a horizontal scaling policy based on performance metrics.

Launch an Amazon MemoryDB cluster and configure the index by using the Hierarchical Navigable Small World (HNSW) algorithm. Configure a vertical scaling policy based on performance metrics.

Launch an Amazon Aurora PostgreSQL cluster and configure the index by using the Inverted File with Flat Compression (IVFFlat) algorithm. Configure the instance class to scale to a larger size when the load increases.

Launch an Amazon DocumentDB cluster that has an IVFFlat index and a high probe value. Configure connections to the cluster as a replica set. Distribute reads to replica instances.

QUESTION 40
A book publishing company wants to build a book recommendation system that uses an AI assistant. The AI assistant will use ML to generate a list of recommended books from the company’s book catalog. The system must suggest books based on conversations with customers.
The company stores the text of the books, customers’ and editors’ reviews of the books, and extracted book metadata in Amazon S3. The system must support low-latency responses and scale efficiently to handle more than 10,000 concurrent users.
Which solution will meet these requirements?

Use Amazon Bedrock Knowledge Bases to generate embeddings. Store the embeddings as a vector store in Amazon OpenSearch Service. Create an AWS Lambda function that queries the knowledge base. Configure Amazon API Gateway to invoke the Lambda function when handling user requests.

Use Amazon Bedrock Knowledge Bases to generate embeddings. Store the embeddings as a vector store in Amazon DynamoDB. Create an AWS Lambda function that queries the knowledge base.
Configure Amazon API Gateway to invoke the Lambda function when handling user requests.

Use Amazon SageMaker AI to deploy a pre-trained model to build a personalized recommendation engine for books. Deploy the model as a SageMaker AI endpoint. Invoke the model endpoint by using Amazon API Gateway.

Create an Amazon Kendra GenAI Enterprise Edition index that uses the S3 connector to index the book catalog data stored in Amazon S3. Configure built-in FAQ in the Kendra index. Develop an AWS Lambda function that queries the Kendra index based on user conversations. Deploy Amazon API Gateway to expose this functionality and invoke the Lambda function.

Option A best meets the requirements because it directly implements a Retrieval Augmented Generation pattern for conversational recommendations using managed Amazon Bedrock capabilities and a scalable vector store. The company’s source data already resides in Amazon S3, which aligns naturally with Amazon Bedrock Knowledge Bases ingestion workflows. A knowledge base can ingest book text, reviews, and metadata, generate embeddings using a supported embedding model, and persist those vectors in a purpose- built vector backend such as Amazon OpenSearch Service. This enables semantic retrieval that is well suited to conversation-driven intent, where user prompts are often descriptive and do not map cleanly to keyword filters.
The requirement to suggest books based on conversations implies the system must interpret natural language context and retrieve relevant passages, reviews, and metadata to ground the recommendation. Knowledge Bases provide managed orchestration for embedding creation and retrieval, which reduces development effort compared to building custom embedding pipelines. OpenSearch Service provides scalable vector search and k- nearest neighbors style similarity retrieval, which supports low-latency responses when properly indexed and sized.
For scaling to more than 10,000 concurrent users, the API layer design in option A is a common AWS pattern: Amazon API Gateway provides a managed front door with throttling and request handling, while AWS Lambda scales horizontally with demand and can invoke the knowledge base retrieval operations. This separates compute scaling from the vector store scaling and helps keep latency predictable under load.
Option B is not the best choice because DynamoDB is not the standard native vector store target for Amazon Bedrock Knowledge Bases in this context and would introduce additional implementation complexity around vector indexing and similarity search behavior. Option C requires substantial ML lifecycle work, model hosting, tuning, and continuous iteration to achieve quality recommendations at scale. Option D provides strong enterprise search, but it focuses on retrieval and FAQs rather than a managed RAG recommendation workflow grounded in embeddings and conversational context for generative responses.

QUESTION 41
A company is creating a workflow to review customer-facing communications before the company sends the communications. The company uses a pre-defined message template to generate the communications and stores the communications in an Amazon S3 bucket. The workflow needs to capture a specific portion from the template and send it to an Amazon Bedrock model. The workflow must store model responses back to the original S3 bucket.
Which solution will meet these requirements?

Create a flow in Amazon Bedrock Flows. Configure S3 action nodes at the beginning and end of the flow to retrieve and store the communications and the model responses. In the middle of the flow, configure an expression to parse each communication. Configure an agent step to send the parsed input to the model for review.

Create an AWS Step Functions Express workflow state machine. Use an Amazon S3 integration GetObject step to retrieve the original communications. Use an intrinsic function Pass step to parse the communications and to pass the results to an Amazon Bedrock InvokeModel step. Configure an Amazon S3 integration PutObject step to store the model responses back to the S3 bucket.

Create an Amazon Bedrock agent that has an action group. Configure instructions to define how the agent should parse the communications. Configure the action group to retrieve the communications from the S3 bucket, invoke the Amazon Bedrock model, and store the model responses back to the S3 bucket.

Create an Amazon Bedrock agent that has a single action group. Configure three AWS Lambda functions in the action group. Configure the functions to retrieve the communications from the S3 bucket, parse the communications and invoke the Amazon Bedrock model, and store the model responses back to the S3 bucket.

QUESTION 42
A company needs a system to automatically generate study materials from multiple content sources. The content sources include document files (PDF files, PowerPoint presentations, and Word documents) and multimedia files (recorded videos). The system must process more than 10,000 content sources daily with peak loads of 500 concurrent uploads. The system must also extract key concepts from document files and multimedia files and create contextually accurate summaries. The generated study materials must support real- time collaboration with version control.
Which solution will meet these requirements?

Use Amazon Bedrock Data Automation (BDA) with AWS Lambda functions to orchestrate document file processing. Use Amazon Bedrock Knowledge Bases to process all multimedia. Store the content in Amazon DocumentDB with replication. Collaborate by using Amazon SNS topic subscriptions. Track changes by using Amazon Bedrock Agents.

Use Amazon Bedrock Data Automation (BDA) with foundation models (FMs) to process document files. Integrate BDA with Amazon Textract for PDF extraction and with Amazon Transcribe for multimedia files. Store the processed content in Amazon S3 with versioning enabled. Store the metadata in Amazon DynamoDB. Collaborate in real time by using AWS AppSync GraphQL subscriptions and DynamoDB.

Use Amazon Bedrock Data Automation (BDA) with Amazon SageMaker AI endpoints to host content extraction and summarization models. Use Amazon Bedrock Guardrails to extract content from all file types. Store document files in Amazon Neptune for time series analysis. Collaborate by using Amazon Bedrock Chat for real-time messaging.

Use Amazon Bedrock Data Automation (BDA) with AWS Lambda functions to process batches of content files. Fine-tune foundation models (FMs) in Amazon Bedrock to classify documents across all content types. Store the processed data in Amazon ElastiCache (Redis OSS) by using Cluster Mode with sharding. Use Prompt management in Amazon Bedrock for version control.

Option B best fulfills all functional, scalability, and collaboration requirements by combining purpose-built AWS services with Amazon Bedrock capabilities. Amazon Bedrock Data Automation is designed to orchestrate large-scale, multimodal data processing pipelines and integrates naturally with foundation models for summarization and concept extraction. Using BDA to process document files ensures consistent preprocessing and model invocation at scale, which is essential for handling more than 10,000 sources per day with high concurrency.
Integrating Amazon Textract for PDFs enables accurate extraction of structured and unstructured text from scanned and digital documents, while Amazon Transcribe is the appropriate service for converting recorded videos into text for downstream semantic analysis. These services are optimized for their respective media types and feed clean, normalized inputs into Bedrock foundation models, improving the quality of contextual summaries.
Storing processed content in Amazon S3 with versioning enabled directly addresses the requirement for version control. S3 versioning provides immutable object history and rollback capabilities without additional complexity. Metadata storage in Amazon DynamoDB supports high-throughput, low-latency access patterns and scales automatically to handle peak upload concurrency.
Real-time collaboration is achieved through AWS AppSync GraphQL subscriptions combined with DynamoDB. AppSync enables real-time updates to connected clients whenever study materials are created or modified, making it well suited for collaborative editing and live synchronization. DynamoDB streams integrate seamlessly with AppSync to propagate changes efficiently.
The other options misuse services or fail to meet key requirements. Amazon SNS does not support collaborative state synchronization, Amazon DocumentDB is not optimized for versioned document storage, Amazon Neptune is unsuitable for document-centric workloads, and Amazon ElastiCache is not designed for durable storage or version control. Option B aligns with AWS best practices for scalable, multimodal generative AI systems built on Amazon Bedrock.

QUESTION 43
A company is designing an API for a generative AI (GenAI) application that uses a foundation model (FM) that is hosted on a managed model service. The API must stream responses to reduce latency, enforce token limits to manage compute resource usage, and implement retry logic to handle model timeouts and partial responses.
Which solution will meet these requirements with the LEAST operational overhead?

Integrate an Amazon API Gateway HTTP API with an AWS Lambda function to invoke Amazon Bedrock. Use Lambda response streaming to stream responses. Enforce token limits within the Lambda function. Implement retry logic for model timeouts by using Lambda and API Gateway timeout configurations.

Connect an Amazon API Gateway HTTP API directly to Amazon Bedrock. Simulate streaming by using client-side polling. Enforce token limits on the frontend. Configure retry behavior by using API Gateway integration settings.

Connect an Amazon API Gateway WebSocket API to an Amazon ECS service that hosts a containerized inference server. Stream responses by using the WebSocket protocol. Enforce token limits within Amazon ECS. Handle model timeouts by using ECS task lifecycle hooks and restart policies.

Integrate an Amazon API Gateway REST API with an AWS Lambda function that invokes Amazon Bedrock. Use Lambda response streaming to stream responses. Enforce token limits within the Lambda function. Implement retry logic by using Lambda and API Gateway timeout configurations.

QUESTION 44
An insurance company uses existing Amazon SageMaker AI infrastructure to support a web-based application that allows customers to predict what their insurance premiums will be. The company stores customer data that is used to train the SageMaker AI model in an Amazon S3 bucket. The dataset is growing rapidly. The company wants a solution to continuously re-train the model. The solution must automatically re-train and re- deploy the model to the application when an employee uploads a new customer data file to the S3 bucket.
Which solution will meet these requirements?

Use AWS Glue to run an ETL job on each uploaded file. Configure the ETL job to use the AWS SDK to invoke the SageMaker AI model endpoint. Use real-time inference with the endpoint to re-deploy the model after it is re-trained on the updated customer dataset.

Create an AWS Lambda function and webhook handlers to generate an event when an employee uploads a new file. Configure SageMaker Pipelines to re-deploy the model after it is re-trained on the updated customer dataset. Use Amazon EventBridge to create an event bus. Set the Lambda function event as the source and SageMaker Pipelines as the target.

Create an AWS Step Functions Express workflow with AWS SDK integrations to retrieve the customer data from the S3 bucket when an employee uploads a new file to the S3 bucket. Use a SageMaker Data Wrangler flow to export the data from the S3 bucket to SageMaker Autopilot. Use the SageMaker Autopilot to re-deploy the model after it has been re-trained on the updated customer dataset.

Create an AWS Step Functions Standard workflow. Configure the first state to call an AWS Lambda function to respond when an employee uploads a new file to the S3 bucket. Use a pipeline in SageMaker Pipelines to re-deploy the model after it has been re-trained on the updated customer dataset. Use the next state in the workflow to run the pipeline when the first state receives a response.

Option D is the best fit because it implements a reliable event-driven MLOps workflow that automates retraining and redeployment with clear orchestration, auditability, and production-grade error handling. The requirement is explicit: whenever a new file is uploaded to Amazon S3, the system must retrain and then redeploy the model used by a web application. A common AWS pattern is to use an S3 event notification to trigger an AWS Lambda function, which then starts a controlled workflow. In option D, Lambda serves as the event handler that reacts immediately to the S3 upload event and passes the necessary context (bucket, object key, dataset version) into an AWS Step Functions Standard state machine.
Step Functions Standard is appropriate for model retraining pipelines because training and deployment steps can be long-running and benefit from durable state, retries, and failure handling. It provides execution history, making it easier to troubleshoot why a particular retraining run failed and to prove which dataset version produced which model version. This operational visibility is critical when the dataset is “growing rapidly” and retraining is frequent.
Within the workflow, Amazon SageMaker Pipelines is the right service to run the ML lifecycle stages in a repeatable way: data processing (if needed), training, evaluation/quality checks, model registration, and deployment to an endpoint used by the application. SageMaker Pipelines is purpose-built for CI/CD-style ML, supporting automated redeployments when a new approved model artifact is produced. By calling a pipeline execution from Step Functions, the company can add governance gates (for example, only deploy if evaluation metrics meet thresholds), and can apply consistent rollback and notification steps when deployment fails.
The other options are weaker: A confuses inference with retraining and does not provide deployment orchestration. B adds unnecessary webhook complexity and describes an awkward event bus configuration. C introduces Autopilot/Data Wrangler, which may be useful but adds extra moving parts and is not required to meet the trigger-and-redeploy requirement.