Back to feed
News Story
APriority75
量子位
1 sources

Xu Mengdi, who returned to China last year, becomes head teacher of Tsinghua's Yao Class

Xu Mengdi, a young scholar who returned from Stanford, joined Tsinghua University's Institute for Interdisciplinary Information Sciences in September 2025 and became one of the new head teachers of the Yao Class (Tsinghua's computer science pilot class). She earned her bachelor's in vehicle engineering from Tsinghua and a PhD in mechanical engineering from CMU, researching robot learning and AI safety. At the orientation, she spoke on 'Building Next-Generation Intelligence,' emphasizing that general-purpose robots need adaptability, not just scaling.

SynthePulse Insight · AI deep readingMembers

From Vehicle Engineering to Yao Class Advisor: Xu Mengdi's Roadmap for 'Next-Generation Intelligence'

Version 1 · 1 source

At the 2026 freshman orientation for Tsinghua's Yao Class, one of the four new class advisors, Xu Mengdi, spoke on the theme of 'Building Next-Generation Intelligence.' How did this young scholar, who studied vehicle engineering as an undergraduate and robotics for her PhD, come to stand at the Yao Class podium? And what does her research trajectory reveal about the Yao Class's new directions?

  • Xu Mengdi joined Tsinghua's Institute for Interdisciplinary Information Sciences in September 2025, becoming one of four freshman advisors for the Yao Class within less than a year of joining.
  • The 2026 Yao Class admitted approximately 90 students, with about 66 entering early through channels such as competition recommendations, and another 24 selected from Tsinghua's new students.
  • Xu Mengdi's doctoral thesis, 'Building Adaptable General-Purpose Robots,' won the Best Doctoral Thesis Award from CMU's Department of Mechanical Engineering. Its core argument is that general-purpose robots cannot rely solely on scaling; they also need adaptability.
Open section navigationThe Unconventional Resume of a Yao Class Advisor

The Unconventional Resume of a Yao Class Advisor

Xu Mengdi graduated from Tsinghua University in 2017 with a bachelor's degree in vehicle engineering. She then studied robotics at Johns Hopkins University and machine learning at Carnegie Mellon University, earning her PhD in mechanical engineering from CMU in 2024. During her doctoral studies, she was advised by Zhao Ding, head of CMU's Safe AI Lab, focusing on robot learning, reinforcement learning, and AI safety.

After completing her PhD, she conducted postdoctoral research at Stanford's Vision and Learning Lab, collaborating with Jiajun Wu and Fei-Fei Li. In September 2025, she officially joined Tsinghua University's Institute for Interdisciplinary Information Sciences as an assistant professor, and within less than a year, she became one of the four freshman advisors for the Yao Class.

Free for now

Read the full analysis

4 more sections of analysis, plus the full takeaway

Loading

Credibility boundary

This article is primarily based on a report from QbitAI, which is a secondary source. Information about Xu Mengdi's personal background, research projects, and awards comes from that report and has not been independently confirmed. Enrollment figures for the Yao Class are based on statistics from the WeChat public account 'Shuimu TsinghuaCent' and are stated as such, not officially confirmed.

Sources for this version

  1. 去年归国的徐梦迪，成了清华姚班班主任量子位

Primary report

量子位

Primary source