Yu Shi

Yu Shi is currently a researcher at Zhongguancun Academy. His research focuses on multimodal scientific foundation models and high-performance machine learning systems. He specializes in joint pretraining of machine learning force fields and structure generation, and multimodal integration with large language models. Additionally, he has long been dedicated to addressing performance bottlenecks in machine learning systems, including low-precision and distributed training strategies, as well as GPU operator design. His work has been published in top-tier machine learning conferences and natural science journals. He has also delivered lectures on Advanced Machine Learning at institutions such as the Institute of Computing Technology, Chinese Academy of Sciences (ICT-CAS), and Tsinghua University. Moreover, he is a long-term contributor to open-source machine learning tools like LightGBM.