Liang Wenfeng
Also known as Wenfeng Liang 梁文锋
Liang Wenfeng founded DeepSeek, the laboratory behind the DeepSeek language model family. He conducted machine vision master’s research with Xiang Zhiyu at Zhejiang University. In a 2023 interview, he described expanding computing resources at High-Flyer and establishing a separate company for large-model research. DeepSeek then released English and Chinese language models. Its 2024 V2 report examined efficiency through selected expert modules and reduced attention memory, while V3 documented further training methods. Liang discussed the team’s research priorities in a 2024 interview. Liang coauthored the 2025 R1 report, which studied reinforcement learning for reasoning.
- Now at
- DeepSeek
- Works
- 26
Publication numbers from OpenAlex, as of 14 October 2026. Citations and h-index are hidden because this OpenAlex record may be incomplete or mix in other authors.
Timeline
-
Around 2010
Machine vision at Zhejiang
Completed machine vision graduate research at Zhejiang University around 2010.
-
2023
A separate research company
Described establishing DeepSeek as a separate large-model research company in 2023.
-
May 2024
DeepSeek-V2 and efficiency
DeepSeek published V2 methods for efficient language models.
-
December 2024
The next model report
The team published the DeepSeek-V3 technical report.
-
January 2025
Training for reasoning
Coauthored the DeepSeek-R1 report with the laboratory’s team.
Sources and credits
Photo: CCTV, via Hangzhou government; Liang Wenfeng is on the right. License: Not stated by source. Source: ehangzhou.gov.cn. Cropped to a square and resized. Ask us to remove or replace this photo.