About

Thanks for stopping by! My name is Ali, and I have 7+ years of experience in research and engineering at the intersection of ML and HCI.

Work

At Amazon, I build AI systems for automated risk detection, such as multi-step tool-calling agents to flag financial fraud. At Apple, I built synthetic datasets and ran evaluations to test upcoming AI/ML features for safety. At Oracle Labs, I worked with the ML Research Group on measuring gender and racial biases in BERT as a function of the model’s scale and pre-training data (arXiv). At Intel, I worked with AI and Analytics team to build ML model explainability and fairness tools—which eventually became an open souce Python toolkit.

Research

My PhD research measured the biased outcomes of machine learning systems used in online advertising. I studied Facebook’s advertising platform to uncover gender and racial biases in the delivery of job and housing ads; disparate pricing in the delivery of political ads; age and gender biases in the delivery of clickbait and deceptive ads.

This work was widely covered in the media (Vox, MIT Technology Review, WIRED), and I had the opportunity to brief staff at the U.S. House of Representatives about our results. Meta later agreed to fix racial biases in their housing ad delivery as part of a settlement with the U.S. Department of Justice.

Personal

I live in Seattle, WA, with my wife and daugther. Previously, I’ve lived in Boston, MA—where I was a part of Khoury College; Saarbrücken, Germany—where I was a student at Saarland University, and a research assistant at the Max Planck Institute for Software Systems; and Lahore, Pakistan, where I am originally from.

I am bilingual in Punjabi and Urdu, and speak just enough German to order a döner kebab. In the rare free time I have, I enjoy reading, single player video games, tabla, and film photography.

Header: Boston’s Back Bay as seen from Longfellow Bridge, shot on Kodak Gold 200.

Last updated: February 26, 2025