SFJ Technologies
Corporate bankruptcy prediction models with 89% accuracy ratio. Automated ETL pipelines for alternative data across 6,000 European securities. End-to-end S&P Global integration serving 22,000 companies weekly.
Quant | AI/ML | Engineer
Building systems at the intersection of mathematics,
software engineering, and financial markets.
Work
Corporate bankruptcy prediction models with 89% accuracy ratio. Automated ETL pipelines for alternative data across 6,000 European securities. End-to-end S&P Global integration serving 22,000 companies weekly.
Dynamic risk-parity optimization using convex methods. Regime-switching frameworks with Hidden Markov Models and Kalman-filtered factors. 15% volatility reduction, 6% Sharpe improvement.
AI explainability pipelines for enrollment forecasting. Interpretable machine learning for 65+ universities, 50,000+ students.
Education
GPA 3.8. Algorithmic trading, derivatives, portfolio construction, stochastic modeling. Teaching Assistant for Machine Learning in Finance. Outstanding Student Award. AWS Innovation Challenge Winner.
GPA 3.6. Stochastic processes, convex optimization, asymptotic statistics. Machine learning, deep learning, algorithms.
Research
Efficient attention mechanisms with Performer architectures. Relative positional encoding variants reducing complexity from O(n²) to O(n).
Hidden Markov Model variants for market regime detection in macroeconomic data. Custom EM and forward-backward implementations.
Craft
Python, C++, C, SQL, UNIX
Bloomberg, S&P Global, AWS
French, Arabic, Spanish, Chinese