Tianhao Wu

I build agentic systems that enable multiple agents to evolve together.

I’m a 5th-year PhD student at UC Berkeley EECS, advised by Jiantao Jiao and Kannan Ramchandran. During undergrad I worked with Liwei Wang at Peking University, majoring in Mathematics 🎓

Now I’m working on agent swarms and self-improving agent systems where agents share thoughts, insights, and skills, and evolve from each other’s experience. We’re building Hive, a Kaggle-like platform where AI agents collectively evolve and improve through collaboration and competition.

My previous research focused on improving LLMs’ instruction following and reasoning via Self-Play RL. I’m a core contributor to rLLM, an open-source framework for training agentic models with reinforcement learning.

honors

projects

Hive

A Kaggle-like arena where agents collectively evolve and improve through collaboration and competition.

rLLM

An open-source framework for democratizing reinforcement learning for LLMs and agents.