Member-only story
Building Your First AI Agent (That Will Actually Improve You As An AI Engineer)
A guide to building a helpful Multi-agent AI system that helps you find the top 10 AI research papers published in ArXiv every day
It’s great to have an in-depth understanding of Agentic AI as an engineer today, and this lesson is all about that.
Let’s begin our journey learning the basics.
Is An Agent Just an LLM?
Most people don’t get this right.
An LLM is a language model that can understand and generate human-like text. It can answer questions, write stories, and help with language translation.
On the other hand, an AI agent is more like a robot that can perform tasks independently.
It has “agency”.
An AI agent can:
- Perform actions and take feedback from its environment
- Think through to improve its actions based on the feedback
- Work towards completing a goal rather than waiting for it to be further prompted (unless specified to do so)
- Work in collaboration with other agents
An agent can be built on top of a large language model (LLM) or other neural networks, and it may use reinforcement learning (RL) or rule-based systems, and so on.
The Basics Of Our Agentic Framework
We start building our AI agents using CrewAI.
It is an open-source Python framework that helps us build AI systems using single or multiple agents that can work together to solve complex tasks.
You might have seen many LangChain tutorials on building agents, but we use CrewAI in this tutorial because LangChain has a peculiar syntax that can get tough to interpret pretty soon.
CrewAI is much simpler to use, is rapidly growing, has an active and helpful community, and has an amazing documentation (surprisingly written by agents built on top of CrewAI).
