With the successful release of GPT-5, the Chinese are still the mainstay. Elaine Ya Le is the leader of the GPT-5 "model system" this time. She said on X: "I am very lucky to be able to lead the team to work hard to make GPT-5 bring a unified user experience to users."
“For the first time, users don’t have to choose between models or even think about model names, enjoying a seamless, unified experience.”
OpenAI realized that having a bunch of models to choose from was a big problem, but they waited until the release of GPT-5 to solve this problem.
This is a huge change in GPT-5 compared to the previous generation, and it is also the big move of "returning to unified intelligence" that Ultraman has been talking about for half a year.
The person in charge of this project is Elaine, which shows her importance in the release and project of GPT-5.
She studied at Peking University with a bachelor's degree in Mathematics and Applied Mathematics. Later, he entered Stanford University in the United States and completed a master's degree in computer science and a Ph.D. in statistics. During his student days, he had experience as a summer analyst at the China Development Bank, a summer deputy at Morgan Stanley, and a research assistant at his alma mater, Stanford.
In early 2024, Elaine joined OpenAI. After getting her Ph.D., she entered the Google system and did applied reinforcement learning research and inspections in the advertising field in Google Brain and other Google teams.Censor modeling and recommendation systems.
Interestingly, some Chinese among the core personnel of GPT-5 mentioned by Elaine have been poached by Meta. This is the case for Hongyu Ren. He was one of the first researchers to be poached from OpenAI to Meta and has joined Meta's latest super-intelligence team.
The gpt-oss open weighted inference large model released two days before GPT-5 is also an important work of Ren Hongyu in OpenAI. He is involved in two such important projects of OpenAI. It is no wonder that Xiao Zha wants to pull him away from OpenAI.
Elaine also specially @several partners. Among them is Tianfu Fu, who is also a core member of GPT-5, but there is not much public information. Currently we only know that he has a PhD from Rice University and only joined OpenAI in January this year. Before that, he was a member of Meta.
In addition, at the GPT-5 press conference, there was also a Chinese who participated in the live speech, Ruochen Wang. He may be one of the most noteworthy Chinese rising stars in OpenAI at present.
Wang Ruochen just graduated with a PhD last year and joined OpenAI in February this year. At the GPT-5 press conference, he represented the multi-modal model team, and the latest voice model of ChatGPT was created by him. In his early years, he briefly studied at Shanghai University of Finance and Economics, and later transferred to the University of Michigan to complete his undergraduate studies. Later, he entered the University of California, Los Angeles to complete his graduate and doctoral studies. His master's degree had to be postponed due to visa issues.
He has also interned at Microsoft and Google, and won the Tomorrow Star Excellence Award for Neural Architecture Search Non-parametric Methods from Microsoft Research Asia. Among interns in the same period, the probability of winning is only one in ten.
It is worth mentioning that Wang Ruochen studied under Professor Cho-Jui Hsieh during his doctoral stage. This professor is a first-class scholar in the academic field, Google Scholar shows that he is cited50,000+, won NSF CAREER, Samsung AI Researcher of the Year, and won the Outstanding Paper Award at ICLR 2021, etc.UCLA officially positions him as aLarge-Scale ML Optimization and RobustnessKey scholars in the field.
Before joining OpenAI, mentors Xie Zhuorui and Wang Ruochen jointly founded TurningPoint AI, a multi-laboratory collaboration focusing on multi-modal agents.
The most eye-catching achievement of TurningPoint AI is that it witnessed the "aha moment" of visual reasoning based on the DeepSeek-R1-Zero method on the 2B model without supervised fine-tuning, which is the first time in the world. Subsequently, this discovery also triggered follow-up in the circle.
Wang Ruochen was one of the authors of this paper that shocked the circle.
GPT-5 core team
In addition to the four important Chinese participants, the core members of the entire GPT-5 team have also surfaced.
Yann Dubois
This Swiss guy joined OpenAI in October 2024. He is one of the leaders of GPT-5 post-training.
He graduated with first place in his undergraduate degree from the Ecole Polytechnique Fédérale de Lausanne in Switzerland. After receiving his master's degree from the University of Cambridge, he interned at universities and research institutions such as Meta, Grab, and the University of Amsterdam, and briefly joined some start-up companies. After receiving his PhD from Stanford in 2024, he joined OpenAI.
At the press conference, he showed off a demo of "Duolingo French Learning Cards" with amazing results.
Christina Kim
She is a veteran in the OpenAI research team. She joined OpenAI as a research scientist in 2020 and has been working there for 5 years.
She witnessed the entire process of Chatgpt from its birth. She is the most typical genius researcher in OpenAI. After graduating from Northwestern University, she co-founded Sourceress, an AI recruitment company, and received $3.5 million in angel round investment from YC. After three years, he left Sourceress to join OpenAI as a research scientist.
Michelle Pokrass
This Russian-born girl is also a member of the OpenAI post-training team. She joined OpenAI in 2022 and has witnessed the entire development process of Chatgpt. Before joining OpenAI, she worked at Clubhouse and Coinbase, once popular applications in Silicon Valley.
Similarly, she only has a bachelor's degree and graduated from the University of Waterloo in Canada in 2018 with a major in computer science.
Eric Mitchell
He only joined OpenAI in 2024 and is responsible for the post-training of the inference model.
He graduated from Princeton University with an undergraduate degree and then studied for a PhD at Stanford University, where he studied under Manning. Before joining OpenAI, he also interned at Samsung Research America, Google, and DeepMind.
Moreover, he is also a member of the Princeton University NACC team and loves music and golf very much.
I have to lament that behind OpenAI’s success is a dense pool of top AI talents. Even though it has been "robbed" by Meta and snatched away at least a dozen researchers, OpenAI still has a rich talent pool.
Looking at their resumes, it is inevitable to speculate: Have they not been targeted by Meta yet, or have they withstood the test of money as reported by Wired not long ago?
No matter what, OpenAI doesn’t dare to gamble. Just the day before GPT-5 was released, Hyperbolic AI CEO Yuchen Jin broke the news on X, saying that a friend working at OpenAI told him that UltramanJust announced a bonus of $1.5 million (over RMB 10 million) per employee over 2 years.