Sunny Qin

I am a fifth-year PhD student at Harvard, advised by Sham Kakade and David Alvarez-Melis, and a member of the Machine Learning Foundations Group.

I work on data-centric machine learning and reinforcement learning. I am broadly interested in the science of foundation models, particularly how natural and synthetic training data interact with learning algorithms to shape model behavior and capabilities.

I'm honored to have been selected as part of the 2025 Apple Scholars in AI/ML.

Blog Posts

RL Excursions thumbnail
RL Excursions during Pre-training: Re-examining Policy Optimization for LLM Training
Backtrack Thumbnail
To Backtrack or Not to Backtrack: When Sequential Search Limits Model Reasoning
Unknowable Thumbnail
Distinguishing the Knowable from the Unknowable with Language Models

Research

Bridging Compute- and Data-Optimal Pretraining
Tian Qin*, Kimia Hamidieh, David Alvarez-Melis.
Preprint [ArXiv]
RL Excursions during Pre-Training: Re-examining Policy Optimization for LLM training
Rachit Bansal, Clara Mohri, Tian Qin*, David Alvarez-Melis, Sham Kakade.
Preprint [ArXiv] [GitHub]
To Backtrack or Not to Backtrack: When Sequential Search Limits Model Reasoning
Tian Qin*, David Alvarez-Melis, Samy Jelassi, Eran Malach.
COLM 2025 [ArXiv]
COMPASS: Benchmarking Constrained Optimization in LLM Agents
Tian Qin, Felix Bai, Ting-Yao Hu, Raviteja Vemulapalli, Hema Swetha Koppula, Zhiyang Xu, Bowen Jin, Mert Cemri, Jiarui Lu, Zirui Wang, Meng Cao.
COLM 2026 [ArXiv] [GitHub]
Distributional Scaling Laws for Emergent Capabilities
Rosie Zhao, Tian Qin*, David Alvarez-Melis, Sham Kakade, Naomi Saphra.
ICML 2026 [ArXiv]
Sometimes I am a Tree: Data Drives Unstable Hierarchical Generalization
Tian Qin*, Naomi Saphra, David Alvarez-Melis.
EMNLP 2025 [ArXiv]
A Label is Worth A Thousand Images in Dataset Distillation
Tian Qin*, Zhiwei Deng, David Alvarez-Melis.
NeurIPS 2024 [ArXiv] [GitHub]
Distinguishing the Knowable from the Unknowable with Language Models
Tian Qin*, Gustaf Ahdritz, Nikhil Vyas, Boaz Barak, Benjamin L. Edelman.
ICML 2024 [ArXiv] [GitHub]
Decomposing Elements of Problem Solving: What "Math" Does RL Teach?
Tian Qin*, Core Francisco Park, Mujin Kwun, Aaron Walsman, Eran Malach, Nikhil Anand, Hidenori Tanaka, David Alvarez-Melis.
Preprint [ArXiv]
Distributional Dataset Distillation with Subtask Decomposition
Tian Qin*, Zhiwei Deng, David Alvarez-Melis.
ICLR Workshops on Data-Centric Machine Learning Research and Data Problems for Foundation Models, 2024 [ArXiv] [GitHub]
Meta-PDE: Learning to Solve PDEs Quickly Without a Mesh
Tian Qin*, Alex Beatson, Deniz Oktay, Nick McGreivy, Ryan P. Adams.
Preprint [ArXiv]

Invited Talks

  • Google DeepMind Frontier AI team, August 2026. "Bridging Compute- and Data-Optimal Pretraining". [Slides]
  • Deep Learning: Classics and Trends (DLCT), October 2025. "Distributional Scaling Laws for Emergent Capabilities". [Event] [Slides]
  • Nvidia, July 2025. "What 'Math' Does RL Teach?" [ArXiv]
  • Seminars on Formal Languages and Neural Networks, March 2025. "Data Drives Unstable Hierarchical Generalization in LMs". [Recording]
  • Voxel 51 Boston AI, ML and Computer Vision Meetup, Feb 2025. "A Label is Worth a Thousand Images in Dataset Distillation". [Event]

Teaching & Service

Personal

Outside research, I enjoy spending my free time outdoors — climbing all sorts of things: rocks, boulders, big wall, frozen waterfalls, and big mountains. I am a board member and an education officer of the Harvard Mountaineering Club. I also recently got into surfing!

Contact

Email: tqin[AT]g.harvard.edu