Build Production-Ready Generative AI Applications on AWS

Why it matters

A retrieval application needs an explicit connection between source documents and vector storage. This exercise places that configuration inside a coffee shop application built with specialized agents.

What to practise

Provision domain-specific data sources and configure Bedrock knowledge bases with S3 sources and OpenSearch Serverless vector storage. Build Strands agents under a Barista Supervisor and test natural-language scenarios. Watch how the knowledge-base configuration connects the source data and vector storage.

How it fits

Focus on 1.4.2’s separation of data sources and knowledge-base storage. The S3 and OpenSearch Serverless configuration gives concrete practice with those resources within the larger agent application.