On the job market

I'm seeking full-time research scientist roles starting early 2026. If there's a good fit, please contact me at {last_name}{first_name[0]}@umich.com.

At a very high-level, my work aims to build verifiable, reliable, and trustworthy reasoning systems. Here's a summary of my relevant work/interests:

I am highly interested in building reliable and universal verifiers for AI reasoning. This includes process reward models for reasoning without step-level human annotations (GRACE) and proposing data-efficient generative process reward models (ThinkPRM) that can scale verifier compute by thinking longer about verification. My work also studied verification bottleneck for self-correction in reasoning.
I am also interested in post-training techniques including aligning LLMs via energy-based models without KL pentalties, a recipe for knowedlge attribution via pretraining data citation and large-scale optimization of model merging to mitigate task tradeoffs.
Most recently, I've focused on evaluation of AI research agents and studying issues with reward design and LLM-as-a-judge.

I am a final-year PhD candidate at the University of Michigan in Ann Arbor advised by Lu Wang and Honglak Lee. I’m currently an intern at LG AI Research working on web agents. I am interested in test-time techniques, reasoning agents, and LLM post-training. I was fortunate to work in different places with so many amazing people. In summer 2024, I was part of CodeGen team at Cohere led by Matthias Gallé where I worked on large-scale model merging. In summer 2023, I was AI2 working with Iz Beltagy and Hao Peng on training models to cite their pretraining data. In 2021, I was Amazon AWS working with Kathy Mckeown. Prior to that, I was an intern at Naver Labs Europe where I worked on Controllable Text Generation and Energy-based models with Hady Elsahar and Marc Dymetman.

Fun fact: I play the piano, write, and produce my own music.

( 🐦 Twitter / 💼 LinkedIn / 🎓 Scholar / 💻 Github / 📄 CV )