Secure Azure LLM Infrastructure

This unit explores how to design, deploy, and secure a three-tier LLM application in Azure using retrieval-augmented generation (RAG) and core platform services. Students will learn how Azure Front Door, App Service, API Management, Storage Accounts, and Cosmos DB fit into an LLM architecture and how to apply secure patterns for data flow, access, and scaling.

Learning Objectives

• Analyze the three-tier LLM application architecture in Azure by mapping components, data flows, and security boundaries between presentation, application, and data tiers
• Apply retrieval-augmented generation (RAG) techniques to build a functional retrieval pipeline that indexes content (using Storage Account or Cosmos DB), performs embeddings, and retrieves context for a sample prompt
• Demonstrate configuring Azure Front Door, Azure App Service, and Azure API Management to securely route, protect, and scale an LLM-backed application (including TLS, basic WAF rules, and rate-limiting policies)
• Evaluate Azure Storage Account versus Cosmos DB for vector and structured data storage in LLM applications and recommend secure storage, access control, and encryption practices aligned with cloud security best practices