About Me
Hi! I'm Qihang Yu, a Ph.D. student at Zhejiang University, advised by Shengyu Zhang and Fei Wu.
My research focuses on LLM Routing, with growing interests in extending routing paradigms to Agent Routing, reinforcement learning for routing optimization, and skill self-evolution in multi-agent systems. I mainly work at the intersection of algorithms and real-world applications, aiming to develop intelligent systems that can dynamically select, coordinate, and improve AI capabilities at scale.
Feel free to reach out if you are interested in collaboration or related discussions.
Research Directions
- LLM Routing Designing adaptive routing strategies to efficiently allocate queries across heterogeneous large language models under constraints such as quality, cost, and latency.
- Agent Routing — Extending routing mechanisms from single-model selection to multi-agent coordination, enabling dynamic assignment of tasks, tools, and workflows.
- RL for Routing — Leveraging reinforcement learning to optimize routing policies and facilitate autonomous skill acquisition and evolution in agent systems.
Education
- 2026 – present · Ph.D. in Artificial Intelligence, Zhejiang University
- 2023 – 2026 · M.S. in Software Engineering, Zhejiang University
- 2019 – 2023 · B.S. in Computer Science, Zhejiang University