Pu (Luke) Yi 易普

Pu (Luke) Yi

Email / Google Scholar / X / CV

About

I am a Ph.D. candidate in computer science at Stanford University, advised by Prof. Sara Achour. My research focuses on efficient machine learning models and systems, particularly those designed for emerging hardware technologies.

Emerging hardware offers opportunities for lower power consumption, nonvolatile storage, and higher storage density. Its computational primitives and error characteristics also create new challenges for algorithms and systems. I am interested in designing learning architectures that fit these technologies, using representations suited to the hardware, resilience to hardware errors, and computation that exploits the physics of the devices.

I enjoy working at the intersection of theory and practice, with broader interests in programming languages, formal methods, software engineering, and mathematics.

Previously, I was a research intern at ByteDance Seed, where I explored efficient attention mechanisms for large language models.

I earned my B.S. in computer science, summa cum laude, from the Turing Class at Peking University. My undergraduate research focused on software testing, advised by Prof. Darko Marinov and Prof. Tao Xie.

Selected Publications

Beyond Research

The characters in my Chinese name, 易普, mean “simple” and “normal,” respectively.

I participated in programming contests in high school, including the Olympiad in Informatics in China, and remained active in the community throughout my undergraduate studies. I served for a year as president of the Peking University Student Algorithm Association (PKUSAA), where we organized programming contests at PKU. I still occasionally help high school students prepare for programming contests.

I enjoy learning languages. I am a native Mandarin speaker, proficient in English, conversational in German and Cantonese, and a beginner in Japanese.