Building LLM Powered Applications

7

Search and Recommendation Engines with LLMs

In the previous chapter, we covered the core steps involved in building conversational applications. We started with a plain vanilla chatbot, then added more complex components, such as memory, non-parametric knowledge, and external tools. All of this was made straightforward with the pre-built components of LangChain, as well as Streamlit for UI rendering. Even though conversational applications are often seen as the “comfort zone” of generative AI and LLMs, those models do embrace a wider spectrum of applications.

In this chapter, we are going to cover how LLMs can enhance recommendation systems, using both embeddings and generative models. We will learn how to create our own recommendation system ...