About Me

Hello! I am Aria.

I’m a researcher in machine learning and mathematics, recently completing dual PhDs at Northeastern University — one in Electrical and Computer Engineering, one in pure Mathematics.

My research is fundamentally about one question: how do we understand and control what neural networks are actually doing inside? That’s led me to work on mechanistic interpretability and activation steering for large language models, game-theoretic methods for explaining feature interactions in black-box models, and adversarial robustness with provable theoretical guarantees.

Along the way I’ve collaborated with clinicians at Brigham and Women’s Hospital on real medical data, built systems that run on frontier LLMs at scale, and published 10+ papers at NeurIPS, ICLR, and AISTATS, including two Spotlight awards.