Azure for Developers: Retrieval-Augmented Generation (RAG) with Azure AI

1h 55mIntermediate2025-01-30

Authors

Ziggy Zulueta

Course details

In this course, Ziggy Zulueta—a Microsoft AI Most Valuable Professional and Certified Trainer—uses examples and practical applications to show you how to leverage Python with Azure Open AI, Cosmos DB, and AI Search to create cutting-edge Retrieval-Augmented Generation (RAG) solutions for enhanced data precision. Dive into RAG fundamentals, Python-based implementations, and performance evaluation methods. Learn how to set up Azure resources, create data indexes, apply skill sets for data enhancement, and automate the indexing process. Explore the importance of vector databases, tokenization, embeddings, and how they facilitate effective data retrieval and augmentation. Evaluate your RAG solutions to ensure accuracy, relevance, and safety. By the end of this course, you will be equipped to develop sophisticated RAG solutions that deliver precise and relevant insights tailored to your business needs.

Learning objectives
Understand what RAG is and what the different components and steps are.
Create a RAG application using Azure AI Search, Azure Cosmos DB and Azure OpenAI.
Evaluate a RAG application using the Azure AI Evaluation SDK.

Skills covered

Azure AI ServicesNatural Language Processing (NLP)Machine LearningGenerative AISoftware Development ToolsArtificial Intelligence (AI)MicrosoftSoftware DevelopmentDeep Dive (X:Y)

Concepts

Introduction

Intro to RAG with Azure AI
Course prerequisites

RAG Core Concepts

Fundamental concepts of RAG
RAG process and architecture
Vector databases
Azure OpenAI embeddings model
Creating a RAG solution with Azure AI Foundry

Azure AI Search

RAG using Azure AI Search
Preparing your resources for RAG
Creating a search index
Creating a data source
Creating a skillset and indexer
Querying your data
Azure AI Search - Import and vectorize data
Sending query results to a language model
Other approaches
Challenge - Create a RAG solution using Azure AI Search
Solution - Create a RAG solution using Azure AI Search

Azure Cosmos DB

RAG using Azure Cosmos DB
Creating Azure Cosmos DB resource
Set up Azure Cosmos DB for RAG
Challenge - Create a RAG solution using Azure Cosmos DB
Solution - Create a RAG solution using Azure Cosmos DB

Evaluating RAG

Evaluation metrics in generative AI
Preparing your evaluation dataset
Evaluate with the Azure AI Evaluation SDK
Challenge - Evaluating a RAG application
Solution - Evaluating a RAG application

Conclusion

Summary and next steps