Microsoft Build Lab 322

Transcript

1. LAB322 | Build, Evaluate and Deploy a RAG-based retail copilot with Azure AI Nitya Narasimhan, PhD & Cedric Vidal Azure AI Advocacy May 21 | 14:15 pm – 15:15 pm PST May 22 | 16:45 pm – 17:45 pm PST May 23 | 15:00 pm – 16:00 pm PST
2. Lab Outline: Welcome to this lab session! This is a self-paced lab. You will use pre-setup machines with the Skillable VM The lab session is 60 minutes. This will cover setup, teardown, and instructor lab review The session is instructor led. Say hello to the instructors – Nitya Narasimhan & Cedric Vidal Call on our proctors if you need help. Say hello to the proctors – Treb & Davinia Sessions are not available on demand. Star the repo & revisit it at home with your subscription
3. Lab Objectives: What Will You Learn In This Session Contoso Chat (v1): A Retail Copilot for Customer Support. What is the application scenario? Retrieval Augmented Generation with Your Data. What is the application architecture? Prompt Engineering to LLM Ops. What is the end-to-end application development lifecycle? Azure AI Studio & Promptflow. What is our development platform for generative AI apps? Contoso Chat (v2): Flex flow, Prompty & AZD. How can I streamline & simplify this further? 1⃣ 2⃣ 3⃣ 4⃣ 5⃣
4. Contoso Chat: A RAG-based Retail Copilot Contoso Web: Frontend A web app for user interactions. It is a client for the chat AI API. Contoso Chat: Backend A generative AI app that takes user query & returns response
5. Architecture: Retrieval-Augmented Generation (RAG)
6. Implementation: Using Promptflow (v1 now – v2 at home)
7. E2E App Lifecycle: Prompt Engineering to LLM Ops
8. Unified Platform: Azure AI Studio
9. Workshop: { Build – Evaluate – Deploy – Use }
10. Related AI Sessions at #MSBuild 2024 Session code Title Date/Time On Demand (Y/N) Breakout Session BRK110 Code-first LLM Ops from Prototype to Production with Gen AI Tools May 22 3:30-4:15pm PST Will be recorded Breakout Session BRK180 Better Together: Build and deploy to Azure with GitHub May 22 11:45-12:30pm PST Will be recorded Breakout Session BRK114 End-to-End AI Development: A Practical Approach May 23 11:00-11:45am PST Will be recorded On Demand OD539 Build Agentic AI Apps with the AutoGen Framework Pre-recorded, Online Session Yes See: https://aka.ms/build/ai for a full list of AI sessions
11. Thank You For Attending This Lab Session ž Stop by the Expert Meetup in the Hub on Level 5 to get your questions answered ž Your feedback is important! Visit https://aka.ms/MicrosoftBuildEvals to complete your session evaluations ž For more information: https://aka.ms/ai-studio/collection
12. © Copyright Microsoft Corporation. All rights reserved. Microsoft branding slide