Hello,

I am Pratik Doshi, an AI focused graduate student at UC Santa Cruz. I hold another Post-Graduate degree in Data Science and earned my undergrad in Finance.

I’m interested in AI research, especially in how models encode complex relationships in the data they are trained on. I’m currently diving deep into Mechanistic Interpretability, a field that seeks to reverse-engineer neural networks with the intention of understanding their inner functioning. I’m currently being advised by Prof. Leilani Gilpin from the AIEA Lab at UCSC. Visit my research notebook for more information about my research.

Over the past year at UCSC, I’ve focused my academic efforts in AI and taken classes like Applied Machine Learning, Deep Learning for Advanced Computer Vision, Artificial Intelligence and Neural Computation. Through these courses, I’ve developed a solid foundation in modern AI.

Through my internship at Data Care LLC (a GPU cloud operator), I’ve developed strong engineering skills in training large AI models in a distributed context. I have also worked as a Software Engineer (both full time and freelance) for over 3 years and have a stronghold on designing, developing and shipping large software systems.