Abstract: In this demo-centric session, we'll start by showing you a generative AI app that brings our own data to an LLM, using Retrieval-Augmented Generation (RAG). We’ll then show you a systematic approach for developing, measuring, and improving generative AI applications, using Microsoft's PromptFlow. We'll evaluate the quality of our current application according to different metrics, we'll make changes to our logic accordingly, and we'll re-evaluate our changes to quantify the improvements made.
Bio: Daniel Schneider is Data Scientist at Microsoft in the AzureML team. He has a background in computer science and machine learning and has worked on different products at Microsoft, including Skype, Windows, and Azure.