I am a Ph.D. student in Pattern Recognition at the Institute of Automation, Chinese Academy of Sciences (CASIA), University of Chinese Academy of Sciences (UCAS). My research interests include large language model post-training, mathematical and multimodal reasoning, tool-using agents, and continual learning.
- 2025–present: Ph.D. student in Pattern Recognition, CASIA / UCAS.
- 2021–2025: B.Eng. in Artificial Intelligence, University of Chinese Academy of Sciences.
- LLM post-training and mathematical reasoning: supervised fine-tuning, reinforcement learning, reasoning data construction, and evaluation.
- Multimodal reasoning and agents: visual tool use, Python-assisted reasoning, and multi-step tool interaction.
- Continual learning: online learning and mitigating catastrophic forgetting.
Python · PyTorch · Transformers · vLLM · verl · Linux · Git · Docker
Experience with SFT, DPO, PPO, GRPO, multi-GPU training, and evaluation on mathematical reasoning benchmarks such as GSM8K, MATH, and AIME.
Email: shihoukun2025@ia.ac.cn