Navigate the generative AI agent tutorial

Use this page to navigate through the generative AI agent tutorial, formerly called the AI cookbook. Follow it from end-to-end, or jump into an area that interests you.

Introduction

Introduction: End-to-end generative AI agent tutorial

10-minute demo

Demo: Mosaic AI Agent Framework & Evaluation

Learn about RAG and AI agents

• Introduction to RAG

• RAG fundamentals

  • Data pipeline

  • RAG chain for inference

  • Evaluation and monitoring

  • Governance and LLMOps

• Improve RAG quality

  • Data pipeline quality

  • Rag chain quality

• Evaluate RAG application quality

  • Define quality

  • Assess performance

  • Enable quality measurement

• Evaluation-driven development

Step-by-step implementation

• Prerequisites: Gather requirements

• Step 1: Clone code and create compute

• Step 2: Deploy a Proof-of-concept to get stakeholder feedback

• Step 3: Curate an evaluation data set

• Step 4: Evaluate the Proof-of-concept quality

• Step 5: Find root causes of quality issues

  • Step 5.1: Debug retrieval quality

  • Step 5.2: Debug generation quality

• Step 6: Iteratively fix and evaluate quality

  • Step 6.1: Fix data pipeline quality

• Step 7: Deploy and monitor the AI application

Next: Introduction to the generative AI agent tutorial >