Building RAG Solutions with Azure AI Foundry (Formerly Azure AI Studio)

Introduction

• Create a RAG solution with little coding

1. Fundamentals of RAG

• The basics of RAG: Adding custom data to your LLM
• Understanding tokens: A key factor of costs in your system
• Vector embeddings: How words connect to each other
• How RAG works: Understanding the process under the hood
• RAG high-level architecture: The required components

2. Introduction to Azure AI Foundry

• Azure AI Foundry overview: Deploy at scale in a safe, secure, and responsible way
• Navigating the Azure AI Foundry
• Creating a project in Azure AI Foundry

3. Setting Up Azure AI Foundry

• Understanding content filters
• Creating content filters
• Creating model deployments
• Navigating the Playground
• Using the Playground and its settings

4. Creating an Index for RAG Using Azure

• Creating an index using Azure AI Foundry
• Creating an index using Azure AI Search
• Understanding retrieval and relevance in Azure AI Search
• Testing your index in the Playground

5. Introduction to Azure Prompt Flow

• Understanding prompt flow
• Create a sample prompt flow for RAG
• Evaluation and monitoring metrics
• Perform evaluations on your RAG system

6. Deploying a RAG Solution

• Deploying the RAG solution using prompt flow
• Testing the REST endpoint using Postman
• Deploying the REST endpoint to Copilot Studio and Microsoft Teams

Conclusion

• Key takeaways
• Additional learning