
Ultimate access to all questions.
No comments yet.
A company is building a generative AI application to provide step-by-step instructions for brewing specialty coffee. The application uses a multi-stage prompt chain that retrieves relevant coffee bean roasting logs from a knowledge base and then generates brewing recommendations. The company needs to ensure low latency, stability, and validated safety controls for production deployment across multiple AWS Regions. Requirements: • Latency must be under 1 second per request, and the company has observed degradation during traffic spikes. • The application must block unsafe or hallucinated recommendations (e.g., dangerous brewing temperatures). • Output consistency for identical inputs must be 99.5% or higher across Regions. Which solution meets these requirements?
A
Use Amazon Bedrock Provisioned Throughput for consistent latency. Implement Amazon Bedrock Guardrails with semantic denial rules to block unsafe content. Use Amazon Bedrock Prompt Management to enforce prompt versioning and approval workflows.
B
Use Amazon SageMaker JumpStart to deploy a custom model with autoscaling. Implement AWS WAF to filter unsafe content. Use Amazon CloudWatch to monitor latency and consistency metrics.
C
Use Amazon Bedrock On-Demand throughput with caching via Amazon ElastiCache. Implement AWS Lambda to validate outputs against a safety database. Use Amazon EventBridge to orchestrate prompt chains.
D
Use Amazon Kendra to improve roast log retrieval accuracy. Store normalized prompt metadata within Amazon DynamoDB. Use AWS Step Functions to orchestrate multi-step prompts.