At Goldman Sachs, I rebuilt a 20+ year-old trading system as a scalable, event-driven architecture with parallel order processing. After eliminating outdated compliance rules and onboarding EMEA users, regional trading activity rose by 50%.
I also built a Python trade reconciliation tool that batch-matches trades against Bloomberg data. It raised the match rate from 94.2% to 99.6% and cut a biweekly manual process from about two hours to minutes.
On Aida, I architected a multi-agent generative AI application with a LangGraph supervisor, deterministic design, and human approval before output ships. My research at Stevens Institute of Technology contributed to UAV classification work that improved model accuracy and was accepted to the 2025 IEEE Aerospace Conference.

