Implementation and Integration I
Explore how to design and implement production-ready Generative AI applications on AWS by integrating services like Amazon Bedrock, AWS Lambda, and Step Functions. Learn to optimize for low latency, streaming responses, human-in-the-loop workflows, and multi-agent orchestration to meet real-world business requirements.
We'll cover the following...
Question 21
A global e-commerce company is launching a “live assistant” chat feature in its mobile app, powered by Amazon Bedrock, to answer questions about orders, returns, and shipping.
The product team requires:
A typing-style experience (token-by-token response streaming).
First-token latency under 300 ms for most requests.
Automatic scaling during traffic spikes (for example, promotions).
Minimal infrastructure management.
End-to-end latency tracing when customers report slow responses.
Which solution best satisfies these requirements?
A. Use Amazon API Gateway (REST API) with AWS Lambda, invoking Amazon Bedrock, and return the full response only after inference completes.
B. Use Amazon API Gateway (WebSocket API) with AWS Lambda that invokes ...