Quantitative Research · Market Microstructure · Statistical Modeling · C++ Systems

Incoming M.S. Financial Mathematics @ University of Chicago
B.S. Computer Science + B.S. Mathematics @ Arizona State University · GPA 4.00

I build quantitative research and trading systems across market microstructure, time-series modeling, Bayesian inference, derivatives, machine learning, and C++.

My work emphasizes reproducible experiments, out-of-sample validation, careful statistical inference, and understanding why models succeed or fail.

📊 BTC/USD Cross-Venue Price Discovery
Market microstructure · synchronized market data · Granger causality · HAC regressions · VAR/VECM

Empirical study of price discovery between Coinbase and Kraken across 10 paired sessions, finding stronger Coinbase-to-Kraken short-horizon predictability across 9 usable sessions.

⚙️ C++ Limit Order Book & Market-Making Simulator
Price-time priority · Nasdaq ITCH replay · Avellaneda-Stoikov · queue diagnostics · performance engineering

Deterministic C++ matching engine benchmarked at ~3.7M events/sec, with multi-seed experiments investigating fill behavior under ITCH-calibrated market flow.

🧪 BayesAudit
Hierarchical LLM oversight · matched experiments · Bayesian monitoring · held-out evaluation

Reproducible empirical study of budget-constrained oversight, including confirmatory and held-out experiments and analysis of a Bayesian-monitor failure mode.

📈 Vol Surface Research Lab
SPY volatility surfaces · SABR/Heston calibration · robustness testing · model diagnostics

Calibration research pipeline for option-chain cleaning, implied-volatility surfaces, static-arbitrage diagnostics, and comparative SABR/Heston model evaluation.

📉 Bayesian Sequential Decision-Making Thesis
HMM filtering · CUSUM regime detection · Bayesian adaptation · adaptive risk control

🔬 Bayesian Market Filters
Kalman/HMM/particle filtering · regime estimation · machine learning · walk-forward evaluation

• Cross-Venue Price Discovery: Found stronger Coinbase-to-Kraken short-horizon predictability across 9 usable BTC/USD sessions using Granger causality, HAC regressions, and VAR-based analysis.
• C++ Market Microstructure: Benchmarked 1M synthetic order events at approximately 3.7M events/sec on Apple M3.
• Market-Making Experiments: Traced ITCH-calibrated fill-rate degradation primarily to sparse executions rather than queue burial.
• Volatility Modeling: Reduced SABR median calibration RMSE from 0.0190 to 0.0077 after deterministic expiry and liquidity filtering while documenting persistent Heston underfit.
• BayesAudit: Found that the confirmatory attacker effect did not replicate in held-out testing and diagnosed a failure mode in the Bayesian/logistic monitor.
• Bayesian Thesis: Reduced regime-detection lag from 15–20 steps to approximately 2 using a volatility-augmented HMM with a CUSUM trigger.
• IMC Prosperity 4: Placed #194 Algorithmic / #256 Overall of 18,800+ teams, finishing in the Top 1.4% overall.

Research Portfolio · Repositories · LinkedIn