Hi, I’m Raj, a CS PhD student at Cornell Tech in NYC. I work on developing responsible, human-centered AI. Advised by Profs. Emma Pierson and Nikhil Garg, my recent focuses have been measuring social discrimination and improving health equity with machine learning.
Previously, I studied CS at MIT with minors in Biology and Women’s & Gender Studies. While there, I worked with Prof. Catherine D’Ignazio and student Harini Suresh at the Data + Feminism Lab. Collaborating with activist groups, we co-designed NLP models to support the difficult labor of tracking gender-based violence (Best Paper, FAccT 2022). The project taught me that naive ML systems often fail at the margins – it takes effort and care to design models for specific, intersectional contexts.
Before that, I explored neural network compression, i.e. improving memory & compute efficiency to mitigate AI’s consumptive footprint. Mentored by Jonathan Frankle, I tested an approach for parallelized pruning of neural networks. During an internship at Apple, I combined compression techniques to rein in the compute footprint of large language models. I also earned Best Paper at BlackboxNLP 2020 for studying how pruning affects interpretability in Transformers.
Though it’s no longer my main interest, I’m also passionate about computational biology, including functional epigenomics and ligand-protein binding prediction. My favorite hobby is cooking, along with other stereotypical grad student activities: lifting weights, baking, reading, and playing tennis. You can find some of my recipes here (it’s a WIP).