Work

FX Arbitrage Engine

Python
WebSockets
Multi-threading
Real-time Systems

Production automated trading system at Argentina's largest retail broker (IOL). Multi-threaded Python engine exploiting FX pricing inefficiencies. Led to promotion to Head of BI.

Trading dashboard

Business Case

The Argentine FX market had persistent pricing inefficiencies due to rapid regulatory changes. Traders were manually executing spread arbitrage but couldn’t keep up with daily policy shifts. I built a multi-threaded Python system that automated the entire process.

Impact

  • Automated the entire trading desk workflow for FX operations
  • Captured ~20% market share through automated pricing reactions
  • Managed institutional capital for the company’s own portfolio
  • Cut churn by 32% using ML-driven user segmentation
  • Reduced client onboarding from 72h to 24h via process automation
  • Led to my promotion from trader to trading desk leadership, then to Head of BI
  • Team: Led a team of 4 at IOL

Technical Details

  • Multi-threaded execution using Python multiprocessing (Pools, Queues, Locks)
  • Real-time price monitoring via WebSocket feeds and broker APIs
  • Automated round-trip trade execution when spreads exceeded thresholds
  • Tkinter GUI for manual overrides by the trading desk
  • Mobile push notifications for critical alerts

View source code on GitHub →