Programming Large Language Models with Azure Open AI: Conversational programming and prompt engineering with LLMs

Chapter 4

Mastering language frameworks

In real life, most of the techniques explored in the previous chapter are not typically implemented from scratch, but rather are used through dedicated frameworks. The most commonly used frameworks are LangChain and Haystack, with Microsoft Semantic Kernel (SK) and Microsoft Guidance gaining ground. Additionally, LlamaIndex (or GPTIndex) is mostly used for the retrieval pipelines to ingest and query data. There are also low-code development platforms, like Microsoft Azure Machine Learning Prompt Flow, for streamlining the flow of prototyping, experimenting, iterating, and deploying LLM and AI applications.

This chapter covers the theory behind and practices for LangChain, Semantic Kernel (SK), and Guidance, ...