A live MLB analytics platform that forecasts player contract ROI and lets users simulate roster moves against real payroll and CBT thresholds projected through 2032. Built with a collaborator, currently serving traffic ahead of a Wall Street Journal feature, with a planned migration off Streamlit to a FastAPI + Next.js stack for scale.

An autonomous intraday equities trading system connected to Interactive Brokers via TWS/IB Gateway, running a cointegration-based pairs strategy with GARCH(1,1) dynamic position sizing, short-selling logic, and a paper-to-live switch. A separate multi-agent layer handles code automation, monitoring, and post-mortem reporting after every session, with the strategy currently live on paper and being extended toward FX, crypto, and futures.

Personal photography portfolio on an interactive 3D globe. EXIF-driven metadata, filterable gallery, journey playback across 200+ photos and 22 places.

A graduate research project (MF821) that fits a VAR(1) cointegration model on intraday mid-prices for five sector-diversified pairs, extracts the cointegration factor via spectral decomposition, and solves free-boundary PDEs by finite differences to derive time-dependent optimal entry/exit bands. Backtested out-of-sample against ad-hoc σ-bands and Bollinger heuristics across 200+ trading days using real NBBO spread costs.

A two-method SOFR yield curve construction project (MF728) benchmarked against Bloomberg’s HW1F price ($217,667.85) for a 1Y×5Y Bermudan receiver swaption. Bootstraps zero rates from OIS swap inputs and fits a cubic spline through Bloomberg’s stripped curve, validated against 6M×6M forwards with <2bp repricing error. Feeds downstream SABR calibration, LMM simulation, and Longstaff–Schwartz pricing.

An LLM-driven sentiment-to-signal framework for equities that scores news and filings with FinBERT, layers in cross-source disagreement features, and converts them to z-score thresholds for entry/exit. Designed as a reproducible out-of-sample evaluation pipeline rather than a curve-fit backtest.

Undergraduate thesis — a four-person team project that built a YOLOv8 object-detection model wrapped in an explainability layer, hitting 92%+ accuracy across medical and environmental datasets. In partnership with Princess Margaret Hospital, applied the pipeline to brain-tumor detection in MRI scans, improving diagnostic accuracy by 18% over baseline.