Accelerate End-to-End Multi-Agents on Databricks and DSPy

Learn to build production-ready GenAI applications using multi-agent systems on the Databricks platform in this 25-minute conference talk. Discover how to create end-to-end workflows that go beyond basic frameworks by leveraging Databricks' unified platform components including Data ETL with DLT and jobs, data storage with Unity Catalog for governance and access control, code development through Notebooks, agent versioning and metric tracking via MLflow and Unity Catalog, evaluation and optimization using Mosaic AI Agent Framework and DSPy, hosting infrastructure with Model Serving and AI Gateway monitoring, and front-end development with Databricks Apps. Explore techniques for building agents that access all your data and models through function calling, then understand how DSPy enables agent-to-agent interaction to ensure accurate question answering. See a practical demonstration of a chatbot powered by multiple agents that can answer questions and reason through specialized topics that base LLMs don't inherently know, providing insights into creating sophisticated AI systems that can handle complex, domain-specific queries through coordinated multi-agent architectures.