Current frame
Algorithm Researcher, ByteDance Seed LLM
Atlas / People / Detail
Jianhui Duan
Jian-Hui Duan is an algorithm researcher in the ByteDance Seed LLM team whose public homepage highlights work on pretraining data, training optimization, and distribution-shift mitigation for large language models.
Profile status: updated
Extended note
Jian-Hui Duan is an algorithm researcher in the ByteDance Seed LLM team. On his public homepage, he describes his work as focusing on pretraining data construction, training algorithms, and data mining for large language models, with broader interests spanning training optimization, distribution shift, and LLM/VLM understanding. The same page lists bachelor's and master's study in computer science at Nanjing University and selected publications in federated learning and large-model research, including the provided Seed technical reports.