Bandit-based framework for automatic variance reduction strategy selection in Monte Carlo option pricing. Two-phase explore-then-exploit with theoretical O(K log N/N) MSE bounds. Tested on 8 option types with near-oracle performance. Includes Bayesian fusion estimator and novel findings on Halton QMC failure in high dimensions.

Modeling cross-asset correlation networks using graph-theoretic methods. Analyzing how correlation structures evolve across market regimes and their implications for portfolio construction and tail risk management.

Python library delivering yfinance's simplicity with better reliability. Automatic multi-source failover to 6 data providers (Yahoo, Stooq, SEC, ECB, Binance, CoinGecko) with per-source circuit breakers, transparent error handling, and 50× concurrent speedup.

Research toolkit for constructing arbitrage-free IV surfaces via SVI parameterization and Heston model calibration. Implements static arbitrage checks, robust Newton-Raphson IV solving, and COS pricing with publication-quality visualizations.