M.S. Student, Columbia University
I am a second-year M.S. student in Computer Science at Columbia University working with Prof. Kathleen McKeown and Prof. Zhou Yu. My research focuses on masked diffusion language models and LLM agents for long-horizon tasks.
I am interested in how language models can revise and reason through iterative infilling, and how agents can learn and operate reliably over realistic long-horizon environments. More broadly, I work on language modeling, reasoning, reinforcement learning, and reliable generative AI.
Before Columbia, I received my B.S. in Computer Science from the University of Minnesota. I was fortunate to work with Prof. Dongyeop Kang at Minnesota and Prof. Jordan Boyd-Graber at the University of Maryland during my undergraduate studies.
Representative work across diffusion language models and long-horizon agents.
EMNLP 2026 Main Conference
Localized diffusion editing for detecting, remasking, and repairing unsupported content without regenerating the full summary.
COLM 2026
Studies reasoning-as-infilling, answer-conditioned reasoning, uncertainty-aware early exit, and adaptive inference in masked diffusion language models.
arXiv, 2026
A benchmark for evaluating computer-use agents on realistic, long-horizon workflows in dynamic environments.
arXiv, 2026
A framework for scalable SFT/RL training of agents inside the same harnesses and environments used at inference time.
EMNLP 2024
Transforms trivia questions into natural information-seeking questions to improve retrieval alignment and cross-domain QA.
arXiv, 2023
A survey of diffusion formulations for text generation, editing, controllability, robustness, and non-autoregressive language modeling.