China looks set to claim another coup in its mission to build AI talent, with one of the world’s top scholars in artificial intelligence for robots listed among PhD supervisors at a Shanghai university.
Su Hao, who holds two doctorates – one in mathematics and one in computer science – has appeared on Fudan University’s faculty list as a professor in electronic information specialising in AI. The list, published on March 31, names 322 PhD supervisors appointed this year.
Previously, Su was a tenured associate professor at the University of California San Diego (UCSD) and the founder and chief technology officer (CTO) of Hillbot, an AI robotics company.
On Google Scholar, his papers on embodied AI are among the most cited globally, with more than 145,000 citations. There has been speculation since October that he was poised to return to China but Fudan’s list is the first confirmation.
According to an interview he gave last year, AI has intrigued him from a young age.
“My earliest interest in AI began in junior middle school, when I came across minimum spanning tree algorithms. That was the first time I felt that human intelligence might not be so unique, or rather, something that machines can reproduce,” he told a tech podcast.
In 2002, Su entered Beihang University in Beijing to study mathematics. In 2005, during an internship at Microsoft Research Asia (MSRA), he had his first real encounter with AI.
After earning his bachelor’s degree, he continued at Beihang to gain a doctorate in applied mathematics.
In 2008, during an exchange at Princeton University in the United States, he served as a student leader and took part in the ImageNet project, led by Li Fei-Fei, who has become a leading figure in AI.
ImageNet, a massive image annotation data set, is considered a cornerstone of the modern deep learning revolution, particularly in computer vision. It structures images according to WordNet’s semantic hierarchy, providing between hundreds and thousands of real images for each visual concept.
“Before ImageNet, people used to think it would take 200 years to crack computer vision. We’ve made amazing progress in just 10 years,” Su said in an October 2024 interview with AsiaTechDaily.
In 2009, he transferred from Princeton to Stanford with Li Fei-Fei, where he pursued a doctorate in computer science to add to the PhD in mathematics he had earned from Beihang.
Su was already a pioneer in 2D and 3D computer vision and embodied AI while still a PhD student, and secured an assistant professor position at UCSD in 2017, before finishing his doctorate.
But he did not stop there. He quickly shifted to robotic AI, now a hot topic in embodied AI.
In 2020, his team released SAPIEN, the world’s first simulation engine capable of modelling interactions between robots and humans. In 2021, he hosted the ManiSkill robot operation competition based on SAPIEN, helping establish evaluation standards for embodied AI.
In July 2025, Su co-founded the AI robotics company Hillbot and served as its CTO, focusing on industrial applications, such as automotive manufacturing, warehousing and retail.
He was a programme chair of the 2025 Computer Vision and Pattern Recognition (CVPR) conference and last year received the CVPR Young Researcher Award and a Frontiers of Science Award, which was awarded in China.
The South China Morning Post has contacted Su for comment.
