Deploying Generative AI and RAG on AWS

This capstone unit covers end-to-end implementation, evaluation, and operationalization of retrieval-augmented generation (RAG) and generative AI solutions on AWS. Students complete hands-on labs and notebooks, analyze case studies, address ethical and governance concerns, and prepare deployment and maintenance plans to remain current with evolving AI capabilities.

Learning Objectives

• Design and implement a retrieval-augmented generation (RAG) solution on AWS that ingests domain data, builds and maintains embeddings, and returns contextually relevant responses in a lab environment.
• Apply custom data pipeline techniques to preprocess, vectorize, index, and update data for scalable RAG workflows and integrate these pipelines with Amazon Bedrock, Amazon Q, and data lake services during hands-on labs.
• Analyze case studies and example notebooks to evaluate model performance, latency, cost trade-offs, common failure modes (e.g., hallucinations), and ethical implications such as privacy and bias, and recommend mitigation strategies.
• Demonstrate operational readiness by developing deployment, monitoring, security, and governance plans for GenAI services on AWS, and outline methods for continual learning and staying current with technical evolution and best practices.