Projects
Trading Systems & Market Microstructure
DeFi Market Making — Hyperliquid
Built short-horizon market-making strategies driven by order-book microstructure, dynamic spreads, and inventory risk control.
- L1/L2/L3 order book data analysis
- Microprice signals and order flow indicators
- Event-time backtesting on 5+ years of tick-level data
- Low-latency execution infrastructure (<100ms)
- Stochastic intensity modeling for limit order book events
CEX–DEX Statistical Arbitrage
Developed a delta-neutral stat-arb system between a centralized exchange and a decentralized derivatives venue.
- Full L2 order book reconstruction
- 2-year tick-level dataset construction
- Microprice deviation signals
- ~8% annualized returns with $800k–$1M daily volume
Avellaneda–Stoikov Market Making
Implemented an Avellaneda–Stoikov-style framework with volatility-adaptive spreads and risk controls.
- Volatility-adaptive spread adjustment
- Inventory-based quote skewing
- Backtested 10M+ trades on DeFi order-book data
- Reduced adverse selection by ~9%
Derivatives Pricing & Quantitative Finance
Multi-Asset Basket Option Pricing
Production-grade pricing engine for multi-asset derivatives combining analytical methods and Monte Carlo simulation.
- Analytical moment-matching (Brigo et al.)
- Monte Carlo with control variate variance reduction
- Term structure modeling with full correlation matrices
- Real market data integration (ECB €STR, Bloomberg)
Trinomial Tree Option Pricer
Built a trinomial tree engine for European & American options with early exercise logic and Greeks.
- European and American option pricing
- Greeks computation (Delta, Gamma, Theta, Vega, Rho)
- Black–Scholes convergence validation
- Lightweight API for pricing and diagnostics
GDP Forecasting with MIDAS
Replicated and extended MIDAS regressions for GDP nowcasting using mixed-frequency financial data.
- MIDAS models with separate lag/lead dynamics
- Daily financial data to forecast quarterly GDP
- Out-of-sample evaluation
- Research-grade implementation
Machine Learning & Data Science
LSTM Stock Price Forecaster
Built a deep learning model for short-horizon financial time-series forecasting.
- LSTM architecture for sequential data
- Feature engineering from historical price and volume
- Technical indicators integration
- Out-of-sample performance evaluation
StackOverflow Tag Prediction
End-to-end machine learning pipeline for multi-label text classification.
- Natural language processing
- Multi-label classification
- Feature extraction and engineering
- Production-ready pipeline design
Infrastructure & Data Engineering
Market Data Aggregation System
Built a comprehensive market data aggregation system with paper trading capabilities.
- Real-time data aggregation from multiple sources
- Paper trading engine for strategy testing
- Data storage and retrieval infrastructure
- WebSocket and REST API integration
DevOps Fullstack Deployment
Full-stack application deployment with modern DevOps practices.
- Docker containerization
- CI/CD pipeline implementation
- Infrastructure as Code
- Cloud deployment