Projects

Production Trading Systems

DeFi Market Making — Hyperliquid (MYR)

Aug 2024 – Jul 2025 | Python

Active market maker on Hyperliquid (DeFi LBO), developing proprietary quantitative strategies in digital asset markets.

Key Features:

• Built short-horizon market-making strategies using L1/L2/L3 order book data, microprice signals, order book imbalance and FIFO queue dynamics
• Developed microstructure-driven alpha signals from depth imbalance, order flow autocorrelation and spread dynamics
• Implemented inventory risk control (Avellaneda–Stoikov inspired) with <100ms latency
• Modeled limit order book event arrivals using stochastic intensity frameworks (Poisson-type flows)
• Event-time backtesting on 5+ years of tick-level data across 12 crypto assets

Technologies: Python, WebSocket APIs, REST APIs, NumPy, Pandas

GitHub: View Repository

CEX–DEX Delta-Neutral Statistical Arbitrage (La Valériane)

Sep 2023 – Jan 2024 | Python

Designed and built an end-to-end delta-neutral statistical arbitrage strategy between Binance (CEX) and dYdX (DeFi DEX), targeting cross-venue microstructure inefficiencies.

Key Features:

• Reconstructed full L2 order books and built 2-year tick-level datasets
• Developed arbitrage signals based on microprice deviations and depth-adjusted fair value estimators
• Implemented slippage modeling and execution constraints
• Achieved ~8% annualized returns with $800k–$1M daily volume
• Maintained strict market neutrality throughout operations

Technologies: Python, Binance API, dYdX API, NumPy, Pandas

GitHub: Cross_EMA_Crypto_Trading_Bot

Academic & Research Projects

Multi-Asset Basket Option Pricing Engine

Personal Project | C#

Production-grade pricing engine for multi-asset derivatives combining analytical methods and Monte Carlo simulation.

Key Features:

• Analytical moment-matching (Brigo et al.) combined with Monte Carlo simulation
• Control variate variance reduction techniques for improved accuracy
• Term structure modeling with full correlation matrices
• Real market data integration (ECB €STR, Bloomberg volatility surfaces)
• Supports European and exotic multi-asset options

Technologies: C#, Monte Carlo methods, numerical optimization

GitHub: Basket_Option_Pricing_Engine_CSharp

GDP Forecasting with MIDAS Regressions

Nov 2025 – Jan 2026 | Paris Dauphine University | Python

Replicated and extended MIDAS (Mixed Data Sampling) regressions for GDP nowcasting using mixed-frequency financial data.

Key Features:

• Implemented MIDAS models with separate lag/lead dynamics
• Used daily financial data to forecast quarterly GDP
• Research-grade Python implementation
• Out-of-sample evaluation and performance benchmarking
• Lag/lead structure optimization

Technologies: Python, Statsmodels, NumPy, Pandas, Time Series Analysis

GitHub: GDP_Forecasting_With_MIDAS_Regressions

Trinomial Tree Option Pricer

Sep 2025 – Dec 2025 | Paris Dauphine University | Python

Implemented a trinomial-tree pricing engine for European and American options with comprehensive features.

Key Features:

• Trinomial tree construction for option pricing
• Support for European and American options
• Early-exercise handling for American options
• Greeks computation (Delta, Gamma, Theta, Vega, Rho)
• Validated against Black–Scholes convergence benchmarks
• Lightweight API for pricing and diagnostics

Technologies: Python, NumPy, SciPy, Financial Mathematics

Algorithmic Trading Systems — Avellaneda–Stoikov Framework

Jan 2024 – Jun 2024 | EPF Capstone Project | Python

Led a team of 6 to develop a comprehensive market-making framework based on the Avellaneda–Stoikov model.

Key Features:

• Implemented Avellaneda–Stoikov market-making framework
• Volatility-adaptive spreads and dynamic inventory management
• Risk controls to limit exposure and adverse selection
• Backtested 10M+ trades on Hyperliquid (DeFi DEXC) order book data
• Reduced adverse selection by ~9%
• Stabilized PnL variance by 7%

Technologies: Python, NumPy, Pandas, Event-time backtesting

GitHub: TRADING_COMPETITION_XTREM_BOT

Machine Learning Projects

LSTM Stock Price Forecaster

Personal Project | Python

Built a deep learning model for short-horizon financial time-series forecasting.

Key Features:

• LSTM architecture for sequential data modeling
• Feature engineering from historical price and volume data
• Technical indicators integration
• Short-horizon forecasting (intraday to daily)
• Performance evaluation on out-of-sample data

Technologies: Python, PyTorch, NumPy, Pandas, Time Series Analysis

StackOverflow Tag Prediction — ML Pipeline

Personal Project | Python

End-to-end machine learning pipeline for multi-label text classification.

Key Features:

• Natural language processing for text data
• Multi-label classification
• Feature extraction and engineering
• Model comparison and evaluation
• Production-ready pipeline design

Technologies: Python, Scikit-learn, NLP, Machine Learning

GitHub: Stackoverflow_Tag_Prediction_ML_Pipeline

Data Engineering & Infrastructure

Market Data Aggregation and Paper Trading System

Personal Project | Python

Built a comprehensive market data aggregation system with paper trading capabilities.

Key Features:

• Real-time market data aggregation from multiple sources
• Paper trading engine for strategy testing
• Data storage and retrieval infrastructure
• WebSocket and REST API integration

Technologies: Python, APIs, Data Engineering

GitHub: Market_Data_Aggregation_And_Paper_Trading

DevOps Fullstack Deployment Project

Personal Project | DevOps

Full-stack application deployment with modern DevOps practices.

Key Features:

• Docker containerization
• CI/CD pipeline implementation
• Infrastructure as Code
• Cloud deployment

Technologies: Docker, Git, Linux, DevOps

GitHub: DevOps_Fullstack_Deployment_Project

Contact

Interested in discussing any of these projects or potential collaborations?

Email: theo.verdelhan@dauphine.eu
LinkedIn: linkedin.com/in/theoverdelhan
GitHub: github.com/theov07