Introduction: End-to-end generative AI agent tutorial
This generative AI agent tutorial (formerly called the AI cookbook) and its sample code take you from a proof-of-concept (POC) to a high-quality production-ready application using Mosaic AI Agent Evaluation and Mosaic AI Agent Framework on the Databricks platform. You can also use the GitHub repository as a template with which to create your own AI applications.
See a list of the pages in the Generative AI agent tutorial.
Tip
There are a few ways you can build a rag app using this tutorial:
- You only have a few minutes and want to see a demo of Mosaic AI Agent Framework & Agent Evaluation.
- You want to get directly into code and deploy a RAG POC using your data.
- You don’t have any data, but want to deploy a sample RAG application.
What do we mean by high-quality AI?
The Databricks generative AI agent tutorial is a how-to guide for building high-quality generative AI applications. High-quality applications are:
- Accurate: They provide correct responses
- Safe: They do not deliver harmful or insecure responses
- Governed: They respect data permissions & access controls and track lineage
This tutorial lays out best-practice development workflow from Databricks for building high-quality RAG apps: evaluation-driven development. It outlines the most relevant ways to increase RAG application quality and provides a comprehensive repository of sample code implementing those techniques.
The Databricks approach to quality
Databricks takes the following approach to AI quality:
- Fast, code-first developer loop to rapidly iterate on quality.
- Make it easy to collect human feedback.
- Provide a framework for rapid and reliable measurement of app quality.
This tutorial is intended for use with the Databricks platform. Specifically:
- Mosaic AI Agent Framework that provides a fast developer workflow with enterprise-ready LLMops & governance.
- Mosaic AI Agent Evaluation that provides reliable, quality measurement using proprietary AI-assisted LLM judges to measure quality metrics that are powered by human feedback collected through an intuitive web-based chat UI.
Code-based workflows
Choose the workflow below that most meets your needs:
Time required | What you’ll build | Link |
---|---|---|
10 minutes | Sample RAG app deployed to web-based chat app that collects feedback | Rag demo |
2 hours | POC RAG app with your data deployed to a chat UI that can collect feedback from your business stakeholders | Build and deploy a POC |
1 hour | Comprehensive quality, cost, and latency evaluation of your POC app | - Evaluate your POC - Identify the root causes of quality issues |