Physics and math student at Pitzer College (Claremont Colleges '28). I split my time between research, building projects, and taking classes across Harvey Mudd, Pomona, and CGU.
Right now I'm focused on two areas: agentic AI systems (multi-agent orchestration, LLM tool use, autonomous decision-making) and geometric methods for machine learning (applying Hilbert space geometry to real-world detection problems). I use Claude Code and multi-model workflows as core development tools.
Geometric Observables for Financial Regime Detection - Solo-authored research paper (3 theorems, 2 propositions) on using projective Hilbert space geometry for unsupervised regime detection. 46-method comparison across 17 historical crises, walk-forward evaluation, Friedman rank testing. Selected poster at the 2026 APS Global Physics Summit. Manuscript in preparation.
MarketMind - Multi-agent autonomous system with 7 AI agent types (rule-based, reinforcement learning via SAC, evolutionary, and LLM-powered). Agents independently observe, plan, coordinate, and adapt. Built with FastAPI, React/TypeScript, WebSocket streaming, and a price-time priority matching engine.
Market Population Dynamics - Agent-based simulation engine written in Rust with PyO3 Python bindings. Three behaviorally-distinct agent populations with replicator dynamics. Maps the full population simplex to find phase boundaries and bifurcation conditions.
Adaptive Volatility Arbitrage - Full-stack options vol arb system. C++ Heston FFT pricer (pybind11), Gaussian Mixture Model regime detection, walk-forward backtester, FastAPI backend, React dashboard. 190+ tests.
I've worked at Perplexity AI (growth) and Quanta Ventures Fund (LLM-based signal generation, quantitative research). I'm comfortable with multivariable (and learning stochastic) calculus, quantum mechanics, and probability.