I recently started my PhD under the supervison of N. Asokan and Florian Kerschbaum. I’m hoping to prevent AI from turning into Skynet.

I worked as a Research Engineer, initially exploring the use of SGX for attesting properties of machine learning models. I later built Amulet because evaluating and comparing multiple machine learning security and privacy defenses was unnecessarily painful.

I completed my Master’s under the supervision of N. Asokan, who taught me how to not suck at writing papers and sharpened my systems security knowledge. I wrote a thesis, On Using Embeddings for Ownership Verification of Graph Neural Networks, that I think turned out pretty neat.

I worked as a Data Scientist at Data Science Dojo, splitting my time between enterprise consulting and teaching data science to industry professionals. Learning how to distill complex ideas for non-technical audiences was an invaluable experience.

I studied Computer Science at Lahore University of Management Sciences (LUMS), where I was first introduced to research through work on internet measurement and censorship analysis with Mobin Javed. I later collaborated with Bimal Viswanath at Virginia Tech on defenses against backdoor attacks in text-based neural networks, leading to my first publication.