LLMs as Autonomous Agents

Session Lead: Mauricio Tec

Session Description: Deploying LLMs as autonomous agents goes beyond text generation––it requires reasoning, planning, and memory. This session explores recent techniques for designing AI agents, including chain-of-thought prompting, memory augmentation, ReAct prompting, and reinforcement learning fine-tuning. Through hands-on coding exercises, participants will gain practical insights into agent design for real-world applications. We will also discuss recent research trends, limitations, and ethical considerations.

Host or Publisher Harvard Data Science Initative