Machine learning systems as tools of oppression

Machine learning systems are increasingly used as tools of oppression. This blog post talks about the harm that ML systems cause and the power imbalance that exists between ML system developers and ML system participants. I also talk about 10 concrete ways for machine learning practitioners to start building fairer ML systems.

I presented a talk version of this blog post on December 3, 2020:

Project Lighthouse — Part 1: P-sensitive k-anonymity

Project Lighthouse is an Airbnb initiative with the goal to reduce discrimination on its platform and to provide leadership for other companies to do the same. This blog post is the first in a 2-part series to explain the methodology underlying Project Lighthouse. This first part gives an overview of the privacy model p-sensitive k-anonymity that Project Lighthouse uses.

How To Hack For Impact: How we started, what we’ve accomplished, and where we’re headed

Hack4Impact is a student-run 501(c)(3) with the goal of inspiring the next generation of technologists, designers, and humanitarians to develop software for social good. Founded at the University of Pennsylvania in 2014, Hack4Impact now has chapters at 11 schools across North America. I was a founding member of Hack4Impact while I was a student at Penn. In this blog post, I reflect upon my experiences joining and then co-leading Hack4Impact at Penn. Specifically, I talk about some of the projects that we worked on and initiatives we launched and how the organization changed and matured over time.