Introduction to AI Orchestration with LangChain and LlamaIndex

Learn how to rapidly build future-proof generative AI apps, locally or in the cloud, using AI orchestration frameworks like LangChain and LlamaIndex.

Introduction

• Building local AI apps with LangChain and LlamaIndex
• What you should know
• Setting up your environment for building AI apps

1. Use AI Orchestration to Build Your First App

• AI orchestration concepts
• Building an app with the OpenAI API
• Running local LLMs
• Your first LangChain app
• Your first LlamaIndex app
• Debugging AI apps

2. Combine LLMs and Indexes to Query Local Documents

• AI over local documents: Retrieval-augmented generation
• Choosing an embedding
• RAG with LlamaIndex
• RAG with LangChain
• Challenge: Document summarization
• Solution: Document summarization

3. Assemble Multi-Step AI Workflows with Chaining

• App concepts for chaining and more complex workflows
• Getting JSON out of the LLM
• LLM function calling
• Challenge: Local LLM task offloading
• Solution: Local LLM task offloading

4. Let the AI Decide What to Do Next: Building Agents

• Introduction to the ReAct agent framework
• Implementing a ReAct agent
• Challenge: LangChain and LlamaIndex strengths and weaknesses
• Solution: LangChain and LlamaIndex strengths and weaknesses

Conclusion

• Next steps for AI app engineers