I have a Master's in Computer Science (Data Science) and a background in discrete math. Most of what I enjoy about this work happens before the modelling: collecting data, figuring out what's actually in it, and getting it into a shape where a question can be answered.
I co-founded an eCommerce company, where I built the data systems we used to make decisions and keep operations from falling over. That's where I learned that the useful answer is usually the one someone can act on, not the one with the best score.
The projects below try to show the process rather than just the results, including the parts that didn't work. Always happy to talk!
A multi-stage pipeline using OLS regression and SVD denoising to forecast weekly sales across 45 Walmart stores.
I cleaned and analyzed housing data, then used XGBoost and ElasticNet to predict home sale prices
Predicting movie review sentiment, then learning a linear map between two embedding spaces to work out which sentences drove the call.
Trying to reproduce CASS, a self-supervised CNN+ViT method, on a brain tumor MRI dataset the authors never used. Some of it held up, some didn't.
A brief Narrative Visualization to examine and explain the relationship between a country's energy usage and income per capita.
An interactive Tableau dashboard visualizing global trends in energy consumption and economic development.