While Chinese artificial intelligence (AI) startup DeepSeek has shocked the global AI industry by introducing low-cost, high-performance models, interest in the “Chinese young geniuses” who led the sensation is hot.
According to Hong Kong’s South China Morning Post and other sources on Monday, the development of the AI model DeepSeek-V3 was led by 150 Chinese researchers and engineers, including DeepSeek founder Liang Wenfeng, and 31 data automation research teams. Researchers at DeepSeek are in their early 20s and 30s. Most of them are from the mainland who graduated from prestigious Chinese universities without any experience studying abroad.
After the launch of the latest model, Luopuri, a developer born in 1995, is drawing the most attention along with Liang Wenfeng. Born in a rural village in Sichuan Province, Luopuri was raised by an electrician father and a teacher mother. Despite his parents’ recommendation to go to a university in the province, Luopuri entered the electronic department of Beijing Normal University, saying, “I have to go to a large city.” By the end of his first year of university, he had transferred to the Department of Computer Science on the advice of a professor who said, “The future of computer science is brighter than that of electronics,” and after graduating from undergraduate, he received a master’s degree in computer language from Beijing University.
Luopuri, recognized as an “AI genius,” received a scouting offer from Xiaomi last month. Xiaomi founder Lei Jun offered Luopuri an annual salary of 10 million yuan, which the industry believes will not accept Xiaomi’s offer, saying it will attract more funds if he starts a business.
Liang Wenfeng is also a native of Korea like Luo Puri. Born in 1985 in Guangdong Province, Liang earned his bachelor’s and master’s degrees in electronic information engineering at prestigious Zhejiang University in engineering. Kaohua and Teng Wan-ding, two key figures in the multi-head potential attention (MLA) study of learning architecture that enhances inference efficiency of deep-seated AI models, are also from the mainland. Kaohua received his degree in physics from Beijing University, while Teng has been studying for his master’s degree at Wuzhen University’s AI lab since 2021. Other key members include Guo Daya, who received his doctorate from Zhongshan University, and Zhu Qihao, who received his doctorate from Beijing University, and Daidamai.
SCMP pointed out that while most Chinese AI startups prefer researchers recognized in the industry or students who have Ph.D.s from abroad, DeepSeek is mainly domestic, which shows that DeepSeek’s approach to talent is different.
SALLY LEE
US ASIA JOURNAL
