1 Developing LLM applications with LangChain

 

This chapter covers

  • Most common LLM applications
  • Understanding LLM application frameworks, their nature, and benefits
  • LangChain's architecture and practical implementation of simple tasks

Interacting with an LLM through a chat interface like ChatGPT is useful for retrieving, summarizing, and generating information. However, for complex tasks, such as enterprise search chatbots or autonomous agents, LLMs are most effective when integrated into software applications. These applications understand natural language queries, retrieve information from sources like vector stores, SQL databases, and REST APIs, and present coherent summaries. They handle unstructured data, import data from diverse sources, and integrate user prompts with LLMs.

LLM applications can become quite complex. Specialized frameworks like LangChain, LlamaIndex, or Semantic Kernel simplify the process. These frameworks incorporate best practices and patterns from various applications, making development more efficient. In this book, you’ll learn how to use the LangChain framework to create LLM-powered applications.

1.1 Building LLM applications

1.1.1 LLM-based engines: summarization and Q&A engines

1.1.2 LLM-based chatbots

1.1.3 LLM-based autonomous agents

1.2 Introducing LangChain

1.2.1 LangChain architecture

1.2.2 LangChain core object model

1.3 Trying out LangChain in a Jupyter Notebook environment

1.3.1 Sentence completion example

1.3.2 Prompt engineering examples

1.3.3 Creating chains and executing them with LCEL

1.4 What You'll Learn from this Book

1.5 Summary