Create Your LLM Agents Leveraging Tools with Unity Catalog Functions

What you’ll learn

AI systems require diverse capabilities to answer business use cases and provide values within each industry.

In this demo, we will build a Compound AI System composed of multiple tools (Unity Catalog Functions) and reasoning agents. A tool typically fetches extra information or executes a computation, such as:

  • Getting the latest customer order status (to answer questions related to an online order)
  • Getting the current weather
  • Calling external Machine Learning model for recommendation
  • Call a Vector Search Index to find relevant information/documents to provide better answers (RAG)
  • Executing specialized routine/LLM queries with specialized prompts or fine-tuned model
  • Execution python script
  • Mathematical operation

Install this demo to discover step-by-step instructions on how to build your first Compound AI System!

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

Recommended

<p>Build High-Quality RAG Apps with Mosaic AI Agent Framework and Agent Evaluation, Model Serving, and Vector Search</p>

Tutorial

Build High-Quality RAG Apps with Mosaic AI Agent Framework and Agent Evaluation, Model Serving, and Vector Search

<p>Mosaic AI: Fine-Tune Your LLM on Databricks for Specialized Tasks and Knowledge</p>

Tutorial

Mosaic AI: Fine-Tune Your LLM on Databricks for Specialized Tasks and Knowledge

Disclaimer: This tutorial leverages features that are currently in private preview. Databricks Private Preview terms apply.
For more details, open the introduction notebook.