Hi!👋 I graduated from Fudan University and am currently building an AI startup in Shanghai, China. I work on LLMs, AI Agents, and AI products.
Since 2026, I have also been a Technical Advisor & Builder at Trooly.AI.
I am a core contributor to the open source projects MOSS and GAOKAO-Bench, focusing on LLM engineering & evaluation.
My current research and engineering interests include agent evaluation, environment and feedback design, agentic role-play, and AI for mathematics.
I finished my research on Chinese Spelling Checking (ACL2023) and multi-turn dialogue (ACL2023) at Fudan University.
By the way, it is worth mentioning that my undergraduate thesis in the Department of Structural Disaster Prevention and Mitigation Engineering at Tongji University pioneered a new direction of using deep learning for structural responses prediction.
- Fudan University, M.S in Computer Science | 2021-2023
- Tongji University, B.S. in Civil Engineering | 2017-2021
- Shanghai AI Lab (2023-2024, Research Intern)
- Zulution.AI / Museland (2024-2025, Algorithm Research Engineer, Early Team Member)
- AI × CAD Exploration (2025-, independent collaboration with researchers from Tongji University)
- Trooly.AI (2026-, Technical Advisor & Builder)
- Moss: An open conversational large language model (MIR)
- Evaluating the performance of large language models on gaokao benchmark (Follow-up work published in ACL2024)
- Investigating Glyph Phonetic Information for Chinese Spell Checking: What Works and What's Next (ACL2023)
- Does correction remain a problem for large language models?
- Sdcl: Self-distillation contrastive learning for chinese spell checking
- Multijugate Dual Learning for Low-Resource Task-Oriented Dialogue System (ACL2023)
- Email: piglaker@email.com or zxt235813@163.com