Deploying Generative AI Apps

This capstone unit guides students through prototyping, securing, and deploying a generative AI web application using Amazon Bedrock and related AWS tooling. Students will learn continuous learning approaches for security and privacy, experiment with AWS playgrounds, perform hands-on prompt engineering, and build a Flask-based movie recognizer application ready for iterative improvement.

Learning Objectives

• Analyze security, privacy, and continuous learning requirements for LLM-powered applications and design an automated pipeline that incorporates monitoring, data labeling, and model updates to mitigate risks
• Apply Amazon Bedrock and AWS playground tools to prototype, test, and compare model behaviors and prompt variations, documenting results and selecting configurations for production
• Demonstrate the ability to build and deploy a Flask-based generative AI web application (movie recognizer) that integrates model inference, user input handling, and basic privacy controls
• Evaluate and refine prompts and model outputs through iterative experimentation and measurable metrics (accuracy, relevance, and privacy compliance), producing a plan for ongoing improvement