I’m a Member of Technical Staff at Mercor, working on code data and applied AI. I build frontier evaluations and reinforcement learning environments for AI labs and enterprises.
I’m interested in language model reasoning, coding agents, verification, and how we craft the data, environments, and feedback that help them improve.
Previously, I worked at Amazon AGI, AWS, and Scale AI, and researched language models and medical robotics at Johns Hopkins. I also directed HopHacks.
My previous work includes QAagent, a multiagent system for unit test generation published as an AAAI 2025 student abstract; SocialNLI, a dialogue-centric social inference dataset; and DeonticBench, a benchmark for reasoning over rules. I also built and evaluated a mobile app for surgical robot teleoperation.
If you want to learn more about me, check out my personal website.
