About me
I am an independent ML researcher studying debate for scalable oversight.
Previously, I was a Research Assistant at Academia Sinica in the lab of Professor Lun-Wei Ku, where I researched domain adaptation in retrieval-augmented generation across model scales and abilities.
Before that, and concurrently for some time (working double time), I was founding engineer of Acaceta, which was a knowledge solutions platform for medical sales representatives and doctors.
I earned my bachelor’s degree from the University of California, Berkeley, where I graduated magna cum laude (3.98/4.00) with a major in Data Science, a concentration in Applied Mathematics and Modeling, and a minor in Chinese. At Berkeley, I wrote an honors thesis, in which I applied natural language processing (NLP) techniques to quantify bias in Taiwanese media. I was advised by Lucy Li.
At the Haas School of Business, I was a member of Angus Hildreth’s Social Psychology and Business Lab for four semesters (Spring 2022-Spring 2024), where I conducted experiments relating to organizational behavior.
I speak English and Mandarin fluently. I studied abroad in Taiwan in Spring 2023 and returned to Taiwan from 2024-2025 to work at Academia Sinica.
I also speak Japanese at an intermediate level (well enough to navigate life in Japan, but not well enough to work there).
In high school, I was ranked top-50 in Lincoln-Douglas debate worldwide, and during my freshman and sophomore years of college, I coached for DebateDrills, a championship-level debate team.
For fun, I like to read, road bike, cook, learn languages, surf, travel, and play pick-up sports.
