welcome

Multi-agent systems and the LLM agents that comprise them are some of the most discussed and worked on topics in AI today. Everyday tasks like searching the web, as well as more complex tasks like building entire codebases of software applications, are some examples where LLM agents have already been deployed. While they are by no means perfect systems, many recognize the great potential for agent-to-agent interactions to reshape the way many tasks are done in our society today.

There are more than a handful of LLM agent frameworks that exist today, which I have used extensively and even contributed to building. In fact, assuming some prior familiarity with LLMs and having past experiences programming with Python or JavaScript/TypeScript, you can probably already create LLM agents as well as multi-agent systems and have them autonomously perform tasks with any one of these frameworks.

This book, however, employs a hands-on approach to help you gain a deeper understanding of the inner workings of a multi-agent system and the LLM agents that comprise them by having you build these from scratch. This is not so different from how you might deepen your knowledge in LLMs by learning how to implement attention and transformers yourself.