Agentic Systems: Deploy and Evaluate RAG Apps with Databricks AI

Databricks lets you create powerful AI Agents using foundation LLMs, Retrieval Augmented Generation (RAG), Vector Search, PDF extraction, and Mosaic AI Agent Evaluation. With RAG, you can enrich prompts with domain-specific knowledge to deliver smarter, more accurate answers without fine-tuning your own models.

In this demo, you'll learn how to:

• Build tools and save them as Unity Catalog functions

• Create and deploy your first agent with LangChain

• Evaluate your agent and build an evaluation loop to ensure new versions perform better on your dataset

• Prepare documents and build a knowledge base with Vector Search

• Deploy a real-time Q&A chatbot using RAG

• Evaluate performance with Mosaic AI Agent Evaluation and MLflow 3.0

• Scan and extract information using Databricks' built-in ai_parse_document function

• Monitor live agents and review production behavior

• Deploy a chatbot front-end with the Lakehouse Application

To run the demo, get a free Databricks workspace and execute the following two commands in a Python notebook: