C++ · simulation

Market Making & Adverse Selection Strategy

• Designed a C++ limit order book market maker providing continuous two-sided liquidity around fair value, simulating fills and spread capture.
• Implemented inventory-aware quoting with volatility-adaptive spreads, skewing prices and sizes using microprice and order-flow-imbalance signals to control directional exposure and inventory risk.
• Benchmarked fixed, inventory-aware, and signal-driven quoting across 100,000+ simulated fills, using 10-second markouts and inventory variance to quantify the adverse-selection tradeoff.

Python · listed options

Options Pricing & Volatility Surface Engine

• Programmed a Python derivatives pricing engine with Black-Scholes, binomial tree, and Monte Carlo simulation, cross-validating outputs across US equity options to verify convergence and pricing consistency.
• Constructed implied-volatility surfaces from listed option chains across 10 US equities, mapping strike and maturity to reveal volatility smiles, skew, and term-structure dynamics.
• Simulated delta-hedged short volatility positions over 5 years, comparing daily versus weekly rebalancing under transaction costs to quantify P&L variance and hedging cost tradeoffs.

Python · cross-asset futures

Cross-Asset Futures Relative Value & Statistical Arbitrage

• Researched 40 economically linked futures spreads across equity-index, fixed-income, and commodity markets, applying liquidity filters and cointegration tests where 8 survived Benjamini-Hochberg correction.
• Modeled mean-reverting spreads with Ornstein-Uhlenbeck time-series models, estimated a 20-day median half-life, and walk-forward tested hedge ratios, with rolling OLS outperforming Kalman filtering out of sample.
• Converted statistical spread signals into margin-aware futures positions, filtering trades by half-life and expected convergence, leaving 3 profitable out of sample after costs.

Python · FastAPI · cloud

AI-Powered Systematic Trading & Research Platform

• Architected a production-grade, cloud-deployed Python quantitative platform with scalable FastAPI services, spanning 9 market data and fundamentals providers across the full systematic investing workflow through broker-connected execution.
• Engineered a machine learning pipeline and model library with TradingView integration, turning saved models into a reproducible workbench.
• Developed a portfolio optimization and risk analytics engine spanning 11 methods from Black-Litterman to regime-aware allocation, with Monte Carlo simulation quantifying VaR, drawdown, and volatility.
• Built an algorithmic strategy builder converting signal-driven trading rules into backtests across 8 timeframes with transaction costs and slippage.
• Implemented an automated system executing strategies via Interactive Brokers under configurable risk limits, logging trade-level P&L.
• Orchestrated 8 multi-agent LLM teams, using knowledge graphs and RAG with MCP tool calls to run platform functions and reason over outputs, making the entire quant stack operable through natural language.