Check out my linkedin profile or shoot me an email at email@example.com.
I currently work as a data scientist for the Consumer Financial Protection Bureau. Prior that role, I was a technology & innovation fellow on the Development team there. I am truly excited about both the mission and the technical side of my work. The bureau is tasked with preventing financial exploitation of consumers, and going after the frustrating sort of structural discrimination (redlining, payday lending, misleading loan advertisements made by e.g. for-profit colleges, etc) that keeps traditionally disadvantaged groups from living prosperous happy lives. I get to use my programming and data skills to build tools that inform academics, educate consumers, and help enforce regulations. I’ve worked on an open source metadata catalog, prepared a distributed half-terabyte mortgage database, done modeling involving NLP, worked on automation for a DevOps team, and upgrading the back-end of a mapping tool that helps identify discriminatory lending practices.
I was an Insight Data Science Fellow, in the Spring 2015 NYC session, on my way out of academia. I had an absolutely wonderful experience with the program, and am happy to answer questions. I also had super sharp fellow Fellows so I encourage interested employers to talk the program directors (or me) about hiring.
I was an astrophysicist (PhD U. Chicago ’12) until January, 2015. In that past life, I researched cosmology and galaxy formation, most recently as a prize postdoctoral fellow in computational astrophysics at the University of Maryland. I explored the physical processes that govern the colors, shapes, and sizes of galaxies using numerical simulations. I also did a lot of data analysis and modeling, both of those simulations and of observational data from galaxy surveys.
Older Software projects
- Metadata catalog: my first project at CFPB was to help construct a catalog of the properties of various datasets across the Bureau. The goal was to produce something that would be the data hub of the bureau — the first stop for finding pertinent datasets. To do that we’re building an extension to a python web app called CKAN. The installer and extension are open source, but the tool itself is for internal use.
- FriendYourVote.com: a web-app that I wrote over the course of 3 weeks at the Insight Data Science program. The app accurately (0.8 AUC) predicts voting proclivity for the 10+ million New York registered voters. Prediction is done with regularized logistic regression or random forest, using vote histories and some demographic information. I serve the voter data from a MySQL database hosted on AWS (the app itself is Django/Python; repo).
- ART: the Adaptive mesh Refinement Tree Code (C, Fortran; read permission provided upon request)
An N-body+hydrodynamics code that uses a combination of multi-level particle-mesh and shock-capturing Eulerian methods to simulate the evolution of dark matter and gas, respectively. I was one of three administrators/lead developers for this software.
- Yt (Python; open-source repo, website)
An open-source visualization and data analysis package primarily for astrophysics simulations. I am a contributor to yt, and have membership status in the collaboration for making significant contributions (mostly for prototyping and developing a back-end to ART-IO).
- Google Sky Labs (Perl, SQL, KML; website)
A set of labs and plugins designed around Google Sky. The plugins query the largest galaxy database, and create plots (top left of the figure below) depicting properties of specified objects in the view window. Anecdotal educational benefits of these labs were presented and published.
- Code for DC
- Star formation histories