HomeArticle

Which Chinese AI experts are hidden behind Elon Musk's Grok3?

职场Bonus2025-02-19 21:11
Counting the 11 Chinese people in xAI whose backgrounds are known.

36Kr "Workplace Bonus" (ID: ZhiChangHongLi)

At noon on February 18, Beijing time, at the Grok-3 conference where Musk personally appeared for xAI, Two Chinese researchers sitting in the "C position" attracted attention. For "Old Ma", this live broadcast had some implications of "revenge" against OpenAI - Musk emphasized in various ways during the live broadcast that "Grok-3's 18-month research has caught up with and even surpassed the results of OpenAI's five or six years of work".

"Jimmy Ba, Leading research (Chief Researcher)" "Tony, working on the reason team" - the self-introductions of the two Chinese faces can be described as understated. 

"Workplace Bonus" found through public data search that "Tony" next to Musk (the second from the right in the following picture) is one of the Co-founders of xAI Yuhuai Wu. After graduating with a Ph.D. from the University of Toronto in 2021, Wu Yuhuai conducted postdoctoral research at Stanford University. He was once a researcher at Google AI, and his personal research interest is "building machines that can reason" [1]. 

● Grok-3 Conference

 

● Known Chinese Experts in the xAI Team

It is worth mentioning that during his doctoral studies, Wu Yuhuai interned at OpenAI for 4 months and Google DeepMind for 11 months. Now, he did not stay in any company but chose to start his own business to pursue the AGI he desires. 

● Wu Yuhuai's X Homepage, xAI Co-founder, Focusing on Reasoning

And the person in the second seat on the left of the picture is even more of a leader in the field. Jimmy Ba is "one of the three giants of deep learning". He studied for his master's and doctoral degrees at the University of Toronto and is a student of Geoffrey Hinton, the "godfather" of AI and a Nobel Prize winner in Physics. Not only that, he is also an assistant professor in the Department of Computer Science at the University of Toronto, and is equivalent to Wu Yuhuai's teacher. More interestingly, Jimmy Ba also interned at Google DeepMind. 

At the same time, They are both founding members of the xAI team.

● Jimmy Ba's X Homepage, xAI Research Head, Engaged in Software and Research

According to the report of the Macro Polo Think Tank [2], among the top artificial intelligence talents in the United States, Chinese researchers account for 38%, and American researchers account for 37%. China has surpassed the United States in the cultivation of talents in the field of artificial intelligence and has become the world's largest exporter of AI researchers. 

This is particularly reflected in the xAI team:

· Among the 12 founding team members, 5 are Chinese, accounting for more than 40%; 

· Among the 4 members of the image team, 2 are Chinese;  

· The core research team continues to have outstanding Chinese joining. 

In July 2023, Musk publicly announced the 12 founding members of the xAI team, of which five are Chinese, accounting for more than 40%. In addition to Wu Yuhuai and Jimmy Ba, there are also Greg Yang (xAI Mathematician), Guodong Zhang, Zihang Dai. There is also a Chinese executive responsible for AI hardware and semiconductor research, Xiao Sun.

● xAI Founding Team

From their educational backgrounds, there are graduates from top domestic universities such as Tsinghua University, Peking University, and Zhejiang University, as well as those with further education experiences from world-renowned universities such as Harvard, Yale, and Toronto. Several scientists have also made significant achievements in their respective fields. 

Greg Yang (Yang Ge) is a mathematician with a high enthusiasm for mathematics. He is from Hunan. He started studying in the United States in junior high school. He graduated from the Department of Mathematics at Harvard University with a bachelor's and master's degree. In 2018, he also won the Morgan Prize, the highest honor in the field of undergraduate mathematics. After graduation, he was recommended to enter the Microsoft Research Redmond Institute as a researcher. He previously participated in the cooperation between Microsoft and OpenAI on large models. Now he is focusing on developing a framework for understanding large neural networks, "Tensor Programs". 

● Greg Yang

Guodong Zhang (Zhang Guodong) graduated with a bachelor's degree in Information Engineering from Zhejiang University and a doctoral degree from the University of Toronto, under the tutelage of Roger Grosse, Focusing on training, adjusting, and aligning large language models. He has won the Apple Doctoral Fellowship in 2022 and the First Prize (1.5%) in the National Undergraduate Mathematical Modeling Competition in 2015, etc. 

● Zhang Guodong, xAI Founding Member

Zihang Dai (Dai Zihang) graduated with a bachelor's degree from Tsinghua University and a master's and doctoral degree from Carnegie Mellon University. He interned at NetEase for one year, and then joined Baidu in 2013 as an engineer at Baidu's Institute of Deep Learning. During his postgraduate studies, he also interned at Baidu in the United States. During his doctoral studies, he interned at Google, focusing on language processing. After obtaining his doctoral degree, he worked as a researcher at Google Brain for 4 years before joining xAI. 

● Dai Zihang, xAI Founding Member

Xiao Sun, former researcher at IBM T.J. Watson Research Center and Meta research scientist. Xiao Sun is a Ph.D. from Yale University, under the tutelage of Professor T. P. MA Group, and graduated from Peking University with a bachelor's degree. After graduation, he spent more than six years at IBM, focusing on the research of machine learning hardware and algorithms. He joined xAI in September 2023. 

● Xiao Sun, xAI Executive, Focusing on AI Hardware and Semiconductor Research

Not only are the majority of the founding team Chinese, according to the news from "New Intelligence Element", the text-to-image model Aurora released by Grok adopts the most cutting-edge MoE architecture at present. Its team achieved the construction from 0 to 1 in just 6 months. And the team has only 4 people, of which there are two Chinese: Haotian Liu and Lianmin Zheng (Zheng Lianmin). 

Among them, Haotian Liu graduated with a bachelor's degree from Zhejiang University and a doctoral degree from the University of Wisconsin-Madison in May 2024, under the tutelage of Yong Jae Lee. Haotian Liu is interested in computer vision and machine learning, and participated in the development of Llava, Grok-1.5V and Grok-2. His recent focus is on building feasible large-scale models. 

● Haotian Liu, Member of the Text-to-Image Team

Lianmin Zheng graduated with a bachelor's degree from the ACM Class of Shanghai Jiao Tong University and a doctoral degree from the University of California, Berkeley in 2024. Her advisor is Ion Stoica and Joseph E. Gonzalez. Lianmin Zheng's research interests include machine learning systems, large language models, compilers, and distributed systems. 

● Lianmin Zheng, Member of the Text-to-Image Team

In addition, there are many Chinese researchers in xAI, such as: 

Xiaobin Wu, who has worked at Microsoft, Twitter, Facebook, ByteDance, etc., and graduated with a master's degree from Shanghai Jiao Tong University, has many years of development and management experience as a Backend Engineering/Engineering Manager/Engineering Director. 

● Xiaobin Wu, xAI Technician

Yunlong Liu, who graduated from Tsinghua University and worked at Google for nearly 8 years. He graduated with a Ph.D. from Johns Hopkins University in 2017, majoring in Computer Biophysics. At Google, he was mainly responsible for Core ML complier/Runtine (OpenXLA/PjRt) Large-scale ML systerm. Now at xAI, he is responsible for large-scale AI Infra and LLM Pretraining.

● Yunlong Liu

Pin-Lun (Byron) Hsu, who graduated from National Taiwan University and worked at LinkedIn for nearly 3 years. Byron is currently responsible for the full stack of ML system at xAI. He once led the Kernel optimization of the "Liger-Kernel" project, the fastest-growing open-source project at LinkedIn, and was responsible for the Inference Optimization of the LinkedIn SGLang project.