Code, research, and everything built by hand.
Engineering
Developed an end-to-end ML pipeline to predict harmful algal bloom severity in Western Lake Erie, using XGBoost with SHAP interpretability and Gaussian Process spatial interpolation across 8 NOAA monitoring stations
Developed a weighted composite scoring algorithm in Python to quantify municipal drinking water safety across 10 contaminants (E. coli, lead, nitrate, TTHM, etc.), outputting a 1-5 star rating and identifying the limiting contaminant factor for Erie County water districts.
A data transformation pipeline that anonymizes and scales a brand portfolio intelligence report for external presentation. Built during my private equity work at Alka Oaks Partners to prepare deal materials for founder calls.
Descriptive panel analysis of Scope 1 and Scope 2 greenhouse gas emissions for publicly listed U.S. companies, 2010–2024. The pipeline matches LSEG/Refinitiv ESG disclosures to EPA GHGRP facility-level data and EPA eGRID state electricity emissions factors to produce a firm-year panel with 80,633 observations across 5,040 unique firms.
Built a fully playable video poker game for my Programming in Java course at Columbia. Implements standard poker hand evaluation logic with an OOP architecture and a lightweight HTML/CSS front-end.
Research
Collecting and analyzing Scope 2 emissions data using Python (Matplotlib, NumPy, Pandas) and a London Stock Exchange Group API, studying corporate behavior in the context of differing regulatory and energy environments.
Led a research initiative and literature review in multiphase reactive transport in sandstone to support carbon mineralization via imbibition technology and pycnometers. Created MATLAB scripts to interpret imbibition data sets with varied boundary conditions.