I am a Member of Technical Staff at Axiom Math, where I work on post-training for formal mathematical reasoning agents.
Previously, I studied mathematics and computer science at UC Berkeley. There, I worked on robotics and reinforcement learning research as well as mathematical research in harmonic analysis and PDE.
Axiom Math
Co-led development of post-training pipeline for formal (Lean) mathematical reasoning agents.
FormalizationAgent helped formalize the 2014 bounded-prime-gaps result (H₁ ≤ 246) in Lean.
University of Chicago Math REU
Worked with David Bowman on dispersive PDE (concentration compactness, NLS).
PaperGround States, Concentration Compactness, and Blow-Up Dynamics in the Nonlinear Schrödinger Equation
University of Washington Math REU
Worked with Bobby Wilson on harmonic analysis and Fuglede's conjecture.
PaperSpectrality of Product Sets with a Perturbed Interval Factor, with Johannes van Vliet
Berkeley AI Research
Researched long-horizon robotic manipulation and reinforcement learning in Sergey Levine's RAIL lab.
The Voleon Group
Worked on numerical computing and performance engineering with JAX, XLA, multicore CPUs, and GPUs.
UC Berkeley
Studied mathematics and computer science.
TeachingFour-time teaching assistant for EECS 127 (Convex Optimization); contributed to the course reader.
CoursesHere is a list of some of the courses I took.
Writing
-
"Juice Numbers" and the awkward way GPT reasoning efforts work
An investigation of how GPT reasoning-effort settings map to hidden juice values, affect prompt caching, and may be learned during post-training.