Unsupervised Learning
A/B Testing
Quantitative Analysis
Equities
- Built a Python-based A/B testing analysis platform used to analyze large-scale behavioral data and extract actionable signals from 10,000+ experimental samples
- Used unsupervised pattern discovery to identify non-obvious anomalies and structural deviations across experiments before client impact
- Identified 5 critical integration issues and 3 runtime anomalies, reducing rollout risk by 20% across experiments affecting thousands of users globally
Algorithmic Trading
Risk Management
Real-time Systems
Feature Engineering
- Built electronic trading system processing S&P 500 data with neural networks (294K+ parameters) achieving 60%+ model accuracy and sub-100ms signal generation for real-time execution.
- Conceived and developed production-grade system with risk management, position sizing, and automated model retraining, featuring real-time P&L tracking and order execution via Alpaca API.
- Implemented monitoring stack with Prometheus/Grafana tracking win rates, fill rates, model/system latency metrics across 500+ symbols with 50+ technical indicators.