- Developed a suite of fully-automated algorithmic trading systems using Python for the Rotman Interactive Trader (RIT) simulation platform as part of Babson’s advanced trading course (FIN4505).
- Architected a 2500+ line modular trading engine for the COM5 commodities simulation, integrating quantitative models across five domains: fundamental news-based signals, trend-based forecasting, spot-futures arbitrage, transportation logistics, and crack-spread refinery operations. Each submodel was capable of real-time autonomous trade execution with centralized position and forecast management. Also included a specialized arbitrage engine for geographic price misalignment across crude oil hubs (e.g., Alaska to NYC), factoring in dynamic pipeline and storage costs, lease management, and directional signals.
- Built robust tender offer arbitrage algorithms (~300 lines) that dynamically evaluated and accepted profitable tenders based on market liquidity and VWAP thresholds, liquidating positions with microsecond-tuned aggressive limit orders and fallback market orders for end-of-day risk mitigation.
- Designed a market-making strategy (~600 lines) incorporating adaptive spread improvement, NBBO-based dynamic speed bumps, inventory thresholds, spoofing/channel stuffing detection, and short-term trend forecasting, achieving consistent profitability ($11K–13K range per session).
- Designed a predictive, threaded algorithmic arbitrage engine for paired securities (e.g., CRZY_M vs CRZY_A), dynamically detecting bid-ask crosses with volatility- and order flow-adjusted thresholds, and executing latency-aware, net-zero market orders to capture pricing inefficiencies.
- Emphasized complete automation from signal generation to trade execution, with clear abstraction layers for modeling, execution, and monitoring, supporting real-time decision-making with minimal human intervention.