Ai-Agents at Custom AI Solutions

AI Agents are everywhere. Every day, tools, libraries, new use cases, and new products come out using and leveraging AI Agents. Several frameworks have been developed to make them easy to build, but what happens under the hood? A very popular pattern for building AI Agents is the ReAct pattern, meaning Reasoning and Acting. The idea is to get large language models (LLMs) to reason about a problem in a manner analogous to how humans do, by breaking down the problem into smaller steps, reasoning about each step, using tools, and then acting on the results. Let’s walk through the ReAct pattern and how we can use it to build a simple AI Agent that writes blog posts in Ruby.