About Me

I started in physics. Taught it for years. Loved the part where you take messy observations and find the model underneath. Then I pivoted to computer science, then to school leadership, then to consulting, and now to a PhD in data science. The common thread isn't the field. It's the instinct: look at the data, find the pattern, build something that works.

At the NYC Department of Education, I built a 6-person data team and managed 30 people total. We used attendance, credit accumulation, and cohort tracking to identify at-risk students early. Graduation rates at Vanguard HS went from 68% to 90% in one school year. The program is still running after I left.

Now I'm doing a PhD at the DISCS Lab, working on federated learning. The specific problem: how do you train ML models across multiple hospitals without exposing patient data, and without the model systematically underperforming for minority populations? I built FairSwarmto solve the coalition selection piece. It's on PyPI. First-author IEEE Transactions submission.

I also reverse-engineered my grandmother's cookie recipe using data science. That one's not on PyPI yet.