Build Contoso Chat End-to-End

This repository contains the lessons and labs for a structured workshop to build Contoso Chat (a custom copilot application) end-to-end using Azure AI Studio and Prompt flow.

Pre-Requisites

The following are required for this workshop:

  1. A GitHub account - Create a free account
  2. An Azure subscription - Create a free account
  3. Access to Azure Open AI service - Request access here

Familiarity with the following is desirable:

  1. Microsoft Azure Fundamentals - Refresh your knowledge
  2. Microsoft Azure AI Fundamentals - Refresh your knowledge
  3. Microsoft Azure AI Studio (preview) - Refresh your knowledge
  4. Using Python & Jupyter Notebooks - Refresh your knowledge

Lessons

Module Lesson
Lesson 1: Explore Contoso Chat Sample Introducing: Contoso Chat Application
Lesson 2: Explore Copilot Architecture Introducing: Copilot Architecture
Lesson 3: Understand RAG Pattern Introducing: Retrieval Augmented Generation
Lesson 4: Understand Paradigm Shift to LLM Ops Introducing: LLM Ops
Lesson 5: Explore Azure AI Studio Introducing: Azure AI Studio
Lesson 6: Orchestrating with Prompt flow Introducing: Prompt flow extension

Labs

Module Lab
Ideation: Define your Contoso Chat Application Lab: Provision Azure Resources For Contoso Chat
Implementation: Build your Contoso Chat application Lab: Build & Run your custom copilot with Prompt flow
Iteration: Evaluate your Contoso Chat application Lab: Evaluate your custom copilot with small datasets
Integration: Deploy and use your Contoso Chat application Lab: Deploy your custom copilot API to Azure