Question 1

A travel company is building a complex "Vacation Planner" system using Amazon Bedrock Agents multi-agent collaboration:Flight Specialist Agent: Searches flight databases using a Lambda-based action groupHotel Specialist Agent: Searches hotel inventory using an API-based action groupSupervisor Agent: Interacts with the user, breaks down the "Plan a trip to Paris" request, delegates sub-tasks to the specialist agents, and aggregates their findings into a final itineraryThe developer needs to optimize the system to reduce latency for simple, single-domain queries (like "Find me a hotel in Paris") while maintaining the ability to handle complex multi-step requests (like "Plan a complete 5-day vacation to Paris").Which collaboration mode configuration meets this requirement?

Options :

A : Configure the supervisor agent in Supervisor with routing mode. Simple queries are routed directly to the appropriate collaborator agent (Flight or Hotel Specialist), while complex queries automatically trigger full supervisor mode for orchestration across multiple agents.

B : Configure the supervisor agent in Supervisor mode. The supervisor always breaks down the request, invokes collaborator agents in parallel, and synthesizes the results regardless of query complexity.

C : Deploy all three agents as independent agents without multi-agent collaboration. Use Amazon Lex to parse user intent and route to the appropriate single agent, then use a Lambda function to aggregate results from multiple agent invocations.

D : Use AWS Step Functions to create a Choice state that evaluates query complexity. Route simple queries to individual agents via direct Lambda invocation, and complex queries to the Bedrock supervisor agent.

Answer: A

Question 2

An online gaming platform is launching a GenAI-powered support assistant using multiple Foundation Models (FMs) available in Amazon Bedrock (including Amazon Titan and Anthropic Claude). The security team has issued strict requirements for the assistant's input handling:PII Protection: Users must be prevented from submitting email addresses or phone numbers.Toxicity: Hate speech and insults must be blocked immediately.Custom Blocking: A specific list of known "cheat codes" and "exploit keywords" must be blocked from entering the model context.Consistency: These rules must apply uniformly across all used FMs without rewriting application logic for each model.Which solution should you implement to meet these requirements with the LEAST operational overhead?

Options :

A : Develop a pre-processing AWS Lambda function using the Python re library to detect PII patterns and cheat codes. Use Amazon Comprehend DetectSentiment API to identify hate speech. Configure the application to call this Lambda function and verify the result before calling InvokeModel on Amazon Bedrock.

B : Implement a standardized system prompt across all FMs that instructs the models to refuse to answer if the input contains emails, phone numbers, hate speech, or words from the cheat code list. Use the Amazon Bedrock Converse API to ensure the system prompt is strictly enforced by the model’s reasoning engine.

C : Use AWS WAF (Web Application Firewall) attached to the application load balancer handling the chat traffic. Create custom regex rules to block PII patterns and cheat codes. Enable the AWS Managed Rules for Anonymous IP List to reduce toxicity risks from bad actors.

D : Create an Amazon Bedrock Guardrail. Configure PII filters to block emails and phone numbers, and enable Content filters for hate speech. Use the Word filters feature to upload the list of cheat codes and exploit keywords. Associate the Guardrail with the InvokeModel API calls for all used FMs.

Answer: D

Question 3

A manufacturing company uses a legacy on-premises ERP system to track assembly line issues. The company wants to integrate Amazon Bedrock to automatically analyze "Technician Notes" (unstructured text) and categorize the root cause of failures using the Anthropic Claude 4.5 Sonnet model.The legacy ERP system has a strict limitation: it blocks all user operations while waiting for external API responses and times out if a response takes longer than 300 milliseconds. The GenAI analysis typically takes 3–5 seconds to complete.As a GenAI Developer, which integration architecture should you design to incorporate the Foundation Model (FM) capabilities without disrupting the legacy system's operations?

Options :

A : Use AWS AppSync with a GraphQL subscription. configure the ERP system to subscribe to the RootCauseAnalysis mutation. When a note is saved, trigger the mutation to invoke Amazon Bedrock directly. The subscription connection keeps the channel open without blocking the main ERP thread, allowing the response to stream back when ready.

B : Create an Amazon API Gateway REST API with a synchronous integration to an AWS Lambda function. Configure the ERP to call this API with the notes. Update the Lambda function to increase the memory allocation to 10 GB, ensuring the model inference completes within the 300-millisecond window.

C : Implement an asynchronous, event-driven architecture. Configure the ERP to push the technician notes to an Amazon SQS queue and immediately return success to the user. Use an AWS Lambda function to poll the queue, invoke the Amazon Bedrock model for analysis, and write the classification results back to the ERP database via a separate callback API.

D : Deploy an AWS PrivateLink connection between the on-premises network and the Amazon Bedrock service endpoint. Configure the ERP system to make direct HTTPS calls to the Bedrock InvokeModel API using the private endpoint, bypassing the public internet to reduce network latency below the 300-millisecond threshold.

Answer: C

Question 4

A smart-home device manufacturer wants to use a Large Language Model (LLM) to identify usage trends from millions of daily device logs. Each log contains a specific device ID, the user's exact GPS coordinates, and a timestamped list of voice commands. Privacy regulations require that the manufacturer must not share granular location data or linkable user IDs with the analytics model, but the model still needs to understand regional usage patterns (e.g., "users in Northern Europe use heating more often").Which data pipeline architecture effectively prepares this data for the model while maintaining regulatory compliance?

Options :

A : Ingest the raw logs into an Amazon S3 bucket and use Amazon Athena to create a view that filters out columns containing Device IDs. Point the foundation model to the S3 bucket backing the Athena view and trust the model's system prompt to ignore any remaining location data.

B : Store the raw logs in Amazon S3 and enable S3 Object Lock in compliance mode to prevent modification. Use Amazon SageMaker Data Wrangler to feed the raw data directly into the model for fine-tuning, ensuring that the training job is run within a VPC without internet access.

C : Create an AWS Lambda function that hashes the Device ID and generalizes the GPS coordinates to a city or region level. Write these transformed logs to a separate "Analytics" S3 bucket and grant the foundation model access only to this secondary bucket for ingestion.

D : Implement a sidecar container in the ingestion pipeline that encrypts the GPS coordinates using asymmetric encryption. Pass the encrypted logs to the foundation model and provide the private key only within the RAG retrieval logic so the model can decrypt locations on-the-fly for regional analysis.

Answer: C

Question 5

A financial services institution is building a proprietary 175-billion parameter Foundation Model (FM) to power a real-time fraud detection assistant. The project lifecycle involves two critical phases with distinct infrastructure requirements:Training Phase: The model must be pre-trained on petabytes of encrypted transaction logs. This requires synchronizing gradients across thousands of GPUs with minimal network latency to accelerate convergence.Deployment Phase: The fraud detection assistant must analyze transactions interactively with sub-millisecond latency. The traffic is highly variable, requiring the infrastructure to scale out automatically during market hours and scale in at night.Which combination of architectural strategies should the GenAI Developer implement to meet these requirements?

Options :

A : Deploy the model to Amazon SageMaker Asynchronous Inference. Configure the endpoint to buffer the incoming transactions in an internal queue, allowing the model to process them in large batches to maximize throughput.

B : Deploy the trained model to an Amazon SageMaker Real-time inference endpoint. Configure Target Tracking scaling policies to verify the SageMakerVariantInvocationsPerInstance metric and automatically adjust the number of instances (data parallelism) to handle the fluctuating traffic.

C : Launch the training job on Amazon EC2 UltraClusters (e.g., p4d instances) within Amazon SageMaker. Enable the SageMaker Distributed Data Parallel (SMDDP) library to maximize throughput and efficiently synchronize gradients across the high-speed Elastic Fabric Adapter (EFA) network.

D : Implement SageMaker Model Parallelism (SMP) on a fleet of Amazon EC2 Spot Instances for the inference deployment. Use the lowest-price allocation strategy to ensure the real-time endpoint remains cost-effective during peak market hours.

E : Use Amazon SageMaker Serverless Inference for the training phase. Configure the MaxConcurrency parameter to parallelize the data processing across thousands of ephemeral Lambda-based compute environments to reduce costs.

Answer: B,C

Higher Test Marks with Free Online AIP-C01 Exam Practice

Buying Options: