GitHub

I publish selected research experiments, prototypes, and technical work on GitHub. My current focus is applied machine learning for financial analytics, anomaly detection, and model-driven attribution.

Featured Project

Portfolio Anomaly Detection

A personal research project exploring unsupervised anomaly detection techniques for investment portfolios. The goal is to detect when a portfolio’s behavior becomes unusual relative to market dynamics and historical patterns, and to support investigation into what is driving the divergence.

• Unsupervised anomaly detection using autoencoders and Isolation Forest ensembles
• Market-relative features and cross-sectional signals for divergence detection
• Emphasis on interpretability and attribution (not just detection)
• Designed as an experimentation-friendly research codebase

What’s Next

This page will evolve over time as additional projects are published and documented. For now, the best entry point is the Portfolio Anomaly Detection repository above and the related technical writing linked throughout this site.