Customizing and Integrating FMs on AWS

This unit covers Amazon Bedrock and related AWS services for deploying, customizing, and operationalizing generative AI solutions. Students will learn prompt management, flows, retrieval-augmented generation (RAG/GraphRAG), knowledge-base integration, agents, model evaluation, and guardrails to prepare for secure production deployment in the final unit.

Learning Objectives

• Analyze AWS generative AI services and architecture by comparing Amazon Bedrock, Amazon Q, and supporting services to select appropriate components for given use cases
• Apply Amazon Bedrock features to build and customize foundation models, including prompt management, flows, fine-tuning/customization approaches, and model evaluation using sample datasets or scenarios
• Demonstrate integration of retrieval strategies (RAG and GraphRAG), knowledge bases, and agents within Amazon Bedrock to produce context-aware responses that meet specified user requirements
• Evaluate and implement security controls, guardrails, data automation, and monitoring practices in Amazon Bedrock deployments to address safety, compliance, and operational concerns