HiπŸ‘‹, I'm Chenyu Zhu, a 2027 fall undergraduate at Huazhong University of Science and Technology and currently a member of MAIR Lab. I build models and agentic tools that continually learn and recursively improve their own capabilities. My research spans foundation model post-training, diffusion models, and AI agents. You can find my work in publications, read my thoughts on the blog, find me on GitHub, or reach me by email.
Chenyu Zhu on GitHub

Chenyu Zhu

Undergraduate Student Β· HUST

I explore continual learning and recursive self-improvement in foundation models and AI agents, spanning post-training, diffusion models, and adaptive agentic systems.

I am seeking research collaborations and internship opportunities in Agentic AI and Foundation model training. If you have a collaboration or internship opportunity, feel free to email me.

Chael-Chael on GitHubView profile

Experiences

  • Advised by Prof. Zhiyuan Ma at the School of Computer Science and Technology.
  • Researching improved RL post-training technics and diffusion models.
  • Co-authored I2E for agentic image editing and TMPO for novel diffusion RL.
RL Post-trainingDiffusionAgentic Image EditingVisit

Research Interests

01Continual Learning & RSI

Systems that learn from ongoing experience and recursively improve their own capabilities.

02Foundation Model Post-Training

Improving reasoning, alignment, and adaptability beyond pre-training.

03Diffusion Models & AI Agents

Combining generative models with agents that perceive, reason, and act.

Publications

Open Source Projects

Blogs