People who can run the experiment and train the model

Founding team

• 

  Yingcheng Wu

  Co-founder & CEO

  AI postdoctoral scholar, Stanford · Fudan University

  From a science background, now working on AI for Science at Stanford, on biomedical world models and autonomous labs, with co-first-author papers in Cell, Nature and Science. Focused on bringing real scientific questions, experimental data and verification conditions into next-generation intelligent systems.

  • AI for life science and biomedicine
  • Scientific data and experimental feedback
  • Tumour microenvironment and immune atlases

  Homepage ↗Google Scholar ↗

• 

  Zhenfei Yin

  Co-founder & President

  Postdoctoral researcher, Oxford · PhD, University of Sydney

  Postdoctoral researcher at the University of Oxford and a PhD graduate of the University of Sydney; previously a research fellow at Shanghai AI Laboratory, and part of SenseTime's AGI group before the PhD. Works on foundation models that build AI agents for the physical and virtual world, spanning multimodal reasoning, embodied intelligence, multi-agent systems and AI Scientist systems for automated scientific discovery; representative work includes the open-source multi-agent robotics framework MARS.

  • Multimodal reasoning
  • Embodied intelligence
  • Multi-agent systems
  • AI Scientist systems and automated scientific discovery
  • Open systems and benchmarks

  Homepage ↗Google Scholar ↗

• 

  Ling Yang

  Co-founder & CSO

  Postdoctoral researcher, Princeton · Incoming faculty, Peking University

  Postdoctoral researcher in ECE at Princeton and incoming assistant professor at Peking University. Works on LLM and agent post-training, reinforcement-learning systems, recursive self-improvement and discovery intelligence. Founded the open-source research community Gen-Verse, whose projects have reached tens of millions of cumulative uses across GitHub and Hugging Face; serves as Area Chair for ICLR, ICML and NeurIPS. Representative work includes RLAnything, ReasonFlux, Buffer of Thoughts and MMaDA.

  • LLM and agent post-training
  • RL systems and infrastructure
  • Recursive self-improvement
  • Discovery intelligence

  Homepage ↗Google Scholar ↗