Scholar
Xingjian Shi
Google Scholar ID: P4G6H7oAAAAJ
OpenAI
Citations & Impact
Citations
16,149
H-index
26
i10-index
37
Publications
20
Co-authors
30
list available
Contact
Publications
Resume (English only)
Academic Achievements
- Selected publications include: 'Towards Geospatial Foundation Models via Continual Pretraining' (ICCV, 2023), 'Earthformer: Exploring Space-Time Transformers for Earth System Forecasting' (NeurIPS, 2022), 'Deep learning and the weather forecasting problem -- precipitation nowcasting' (Wiley & Sons, 2021), etc.
Research Experience
- Building multimodal models at OpenAI; Led model development at Boson AI; Senior applied scientist at Amazon, leading AutoGluon Multimodal and DeepEarth projects.
Education
- Obtained Ph.D. from the Hong Kong University of Science and Technology in 2018, supervised by Prof. Dit-Yan Yeung. Received B.E. degree from Shanghai Jiao Tong University in 2014.
Background
- Currently building multimodal models at OpenAI. Before that, led model development at Boson AI, creating the Higgs series of LLM and multimodal LLMs, like Higgs-Llama and Higgs-Audio. Prior to Boson AI, was a senior applied scientist at Amazon, leading two projects: AutoGluon Multimodal and DeepEarth.
Miscellany
- Loves open source and has participated in projects like Apache/MXNet.
