Ph.D. student at Peking University, building research agents, multimodal systems, and post-training pipelines.
- I work on large language models, post-training, reinforcement learning, agent planning, tool use, and multimodal scientific intelligence.
- My recent work sits at the intersection of research automation, scientific content generation, data analytics agents, and benchmark design.
- I care about systems that are both publishable and shippable: strong ideas, clean interfaces, editable outputs, and reproducible evaluation.
- Ph.D. Student in Data Science and Engineering at Peking University
- Algorithm Intern at ByteDance Volcano Engine
- Building agent systems for scientific workflows, data analytics, and multimodal generation
LLM post-training RL and alignment agent planning tool use multimodal systems scientific generation benchmarking
- Email:
zhouliu.919@qq.com - Location:
Beijing - Affiliation:
Peking University