Graduate of Asia's second-best university pushes boundaries of AI-generated video at Goog…

Lijun Yu, a graduate of China's Peking University, went from writing simple computer programs in high school to becoming an award-winning researcher advancing Google's AI video-generation technology.

Yu, who is a senior research scientist at Google DeepMind, was named to MIT Technology Review's 2026 Innovators Under 35 Asia Pacific list for his contributions to AI systems that can understand and generate videos.

Google DeepMind is the company's AI research division, responsible for developing some of its most advanced AI technologies, including Gemini, its family of AI models, and Veo, its video-generation technology.

From school programming to the forefront of AI research

His interest in programming began in high school, when he wrote a computer program to ring a bell at scheduled times after the school's regular bell system was turned off during exams, according to a profile published by Peking University on Oct. 26.

Although his family encouraged him to study economics, he was more interested in computers. He enrolled at Peking University, where he pursued both subjects and graduated in 2019 with bachelor's degrees in computer science and economics. Peking University ranked second in Asia in both the 2026 QS and Times Higher Education regional university rankings.

Yu initially considered specializing in operating systems but turned to artificial intelligence after encouragement from a professor. In 2018, he joined a summer research program at Carnegie Mellon University in the U.S., where he studied how computers analyze videos. He later pursued doctoral research in the U.S. and worked as a research intern at Google, shifting his focus from teaching AI to understand videos to developing systems that could generate them.

His research led to several major projects, including MAGVIT and MAGVIT-v2, which improved how AI processes and generates images and videos.

He also helped develop VideoPoet, a Google AI system that can create videos from written instructions, animate still images and edit existing footage. The research won a Best Paper Award at the 2024 International Conference on Machine Learning, a major international AI research conference.

Yu joined Google DeepMind as a senior research scientist in May 2024, according to his professional CV. He has contributed to several major projects, including Gemini, Veo 3.1 and Nano Banana, an AI tool that allows users to create and edit images.

Despite AI's rapid advances, he believes researchers should focus on problems that genuinely interest them rather than simply following the latest technological trends. He also emphasizes the importance of scientific rigor, arguing that researchers remain responsible for ensuring their findings are reliable, regardless of how much AI assists their work.

For students considering a career in AI, he offers a question: "When the AI buzz eventually fizzles out, would you still want to research it?"