Research Statement

My research operates at the intersection of high-dimensional financial time series, modern applied mathematics, and computational physics. I’ve always gravitated toward non-parametric, structure-seeking techniques—drawing on geometry, spectral methods, and ideas inspired by quantum field theory and statistical mechanics. I view financial markets as constrained dynamical systems with symmetries, coupling structures, and propagation dynamics, not as noisy regression exercises.

I spent over a decade at the University of Chicago, completing Ph.D.-level coursework in economics, mathematics, and physics, including quantum field theory and string theory. That training fundamentally shaped my worldview: I think in terms of conservation laws, constraint surfaces, interacting fields, and renormalization-style flow rather than linear regressions and point estimates. (I’ve always taken crescat scientia vita excolatur literally, and I kept pushing it until UChicago politely hinted that perhaps I had taken the motto a bit too far.)

In parallel, I developed a deep full-stack engineering background. I’ve built distributed systems, low-latency infrastructure, data pipelines, and large-scale portfolio engines; I’ve designed APIs, supervised production rollouts, and managed teams building systems that cannot fail. This engineering fluency informs my research: I don’t think about models in isolation—I think about how they deploy, scale, degrade, and behave under stress. My research is inseparable from the systems that implement it.

My current work brings these threads together: using mathematical and physical priors to build AI models that reason in a financially grounded way, and grounding those models in production-grade compute and software engineering. The goal is to construct systems that internalize balance-sheet mechanics, microstructure, risk propagation, tax/custody constraints, and the causal architecture of price formation—not just token-level correlations. These models must be mechanistic, interpretable, and operationally viable at institutional scale.

Longer term, I aim to build decision architectures that merge practitioner intuition, physics-style structural reasoning, and modern engineering discipline: systems that reason coherently, scale efficiently, and remain robust under real-world stress. This is an attempt to fuse contemporary AI with the mathematical structures and computational rigor required for production-grade portfolio management.