About

Tian (Sunny) Qin is a fourth-year Ph.D. student in Computer Science at Harvard University, co-advised by Professor David Alvarez-Melis and Professor Sham Kakade. Her research aims to build a scientific understanding of how large AI models are trained, with a particular focus on data-centric AI and reinforcement learning. On the data-centric side, she studies how the data used to train models shapes their capabilities and behavior. On the reinforcement learning side, she studies how it can be used to improve the way models reason and solve problems. Across both directions, her goal is to develop a scientific understanding of model training choices in order to build better, more reliable models. Prior to her Ph.D., Sunny earned her B.A. cum laude in Physics with a minor in Statistics and Machine Learning at Princeton University. She then spent several years in quantitative trading before pivoting into AI research and joining Harvard's Ph.D. program. Sunny is also passionate about mentoring others. She has served as a Teaching Fellow for CS 2881: AI Safety, and she mentors students through independent research mentorship of master's and thesis students. She is currently organizing a workshop at NeurIPS 2026 on the interplay between pre- and post-training. Outside of research, she serves as Education Officer of the Harvard Mountaineering Club, where she regularly plans and leads climbing trips and expeditions, teaching and coaching students in climbing and mountaineering skills.

Education

Harvard University

Ph.D. , Computer Science

Class of 2027

Awards & Publications