I am honored to attend and give a talk at the 74th Lindau Nobel Laureate Meeting as a selected young scientist in July 2025.
I am currently a Ph.D. candidate at the College of Chemistry and Molecular Engineering, Peking University, where I also received my B.S. in Chemistry (2018–2022). I am fortunate to be advised by Prof. Luhua Lai and Prof. Zhirong Liu.
I work at the intersection of chemistry and large language models, and I think of myself less as someone who applies AI to chemistry than as an AGI-native researcher. I take it as a working assumption that frontier models will keep absorbing the routine layer of research; what stays scarce is judgment, taste, and the willingness to ask which problems are worth solving at all. Polishing another accuracy point on a saturated benchmark is no longer where I want to spend that judgment.
Concretely, my work falls into three layers:
- Measuring the frontier: building benchmarks and hard evaluations that probe where today's models actually fail on real scientific problems (e.g. SUPERChem, and earlier work mapping the ceiling of single-step retrosynthesis).
- Forging tools: turning model capability into grounded, verifiable tools that scientists actually use, such as SynCraft for synthesizability optimization and Revenant for reaction-failure diagnosis.
- Probing beyond it (AGI+1): using domain insight plus coding agents as leverage to enter problems that would once have taken a whole group years to approach.
Beyond my academic studies at Peking University, my journey has been shaped by continuous engagement with world-class research labs. I have spent extensive time as a research intern within the Microsoft Research (MSR) ecosystem, collaborating with leading experts to push the boundaries of AI for Science. This includes my time at MSRA (May 2022 – Nov 2023), MSR AI4S Beijing (Mar 2025 – Sep 2025), and most recently MSR AI4S Cambridge (Jan 2026 – May 2026), where I explored the intersection of LLMs and chemistry.
To ensure my computational work remains grounded in physical reality, I also served as a visiting research student at the Center for Algorithmic and Roboticized Synthesis, IBS (Oct 2024 – Dec 2024). Advised by Prof. Bartosz Grzybowski, I gained valuable insights into chemical automation and catalyst development, effectively balancing my AI expertise with deep chemical domain knowledge.
I am always open to discussions on AI, Chemistry, and the future of automated science, you can also find my publications on Google Scholar. Feel free to reach out me at catalystforyou@outlook.com!