Agents
In the previous chapter, we explored how large language models (LLMs) leverage tools to enhance their responses and expand their capabilities. This chapter delves into the fundamental concept of an agent within the domain of generative AI applications. Agents possess the ability not only to utilize tools but also to strategically plan which tools to employ, the order of their use, and how many times they should be applied.
We will approach agent construction from two distinct perspectives: first, by using LangChain and LangGraph on Google Cloud with the Gemini SDK, where we will explain how to build agents on the Google Cloud platform and provide illustrative implementation examples. Second, we will briefly touch on the use of managed services like Vertex AI Agent Builder to create agents that are ready for deployment.
We will cover the following main topics in this chapter:
- What is an agent?
- Agents in LangChain
- Vertex AI Agent Builder
