I'm driven by curiosity, with more questions than time to explore them. Most of my time goes into understanding how AI systems learn, remember, and reason, with a growing interest in how they can help us make scientific discoveries.

I'm an MS candidate in Natural Language Processing at UC Santa Cruz, where I work with Dr. Leilani Gilpin and Dr. Chenguang Wang. I received my undergraduate degree from the University of Arizona and was a Research Scholar at the University of Edinburgh, working on neuro-symbolic AI.

My research focuses on reinforcement learning, post-training, memory, and verification for long-horizon agents. I study how models learn from feedback, when their reasoning fails, what they should remember, and how to design training signals they cannot easily exploit. I'm also interested in continual learning and evaluations that predict how systems will behave as they scale.

Alongside my research, I contribute to open-source tools and benchmarks, including rLLM, Terminal-Bench Science, and ARC-AGI-3. These contributions let me explore questions about learning and reasoning through practical work on training and evaluating AI systems.

I'm increasingly drawn to AI for scientific discovery: systems that can challenge their own assumptions, revise beliefs when evidence contradicts them, and help solve consequential problems in biology, medicine, and beyond.

Outside my main research, I like learning from first principles and building things to understand them. Right now, I'm reading about Peter Turchin's work on cliodynamics and re-implementing core ideas behind AlphaFold. My curiosity takes me from historical change to how proteins fold, and how computation can help us understand both.

I want to build AI systems that help us discover things we couldn't discover alone while working on problems that matter to society.