At American Express, I lead development of a transformer-based model that identifies customer intent from sequential website clickstream activity, contributing from research through model development.
I previously built an enterprise control that blocked unprofitable customers from welcome offers, driving approximately $25MM in upfront investment savings. I also evaluated a gaming control that enabled approximately $44MM in cost savings.
My work combines SQL, Python, machine learning, experimentation, and data storytelling. I also built t-SNE analyses of vesicle networks and contributed to a gold-medal iGEM project developing a CRISPR Cas12-based HPV diagnostic kit.

