Summary
Casey Meehan is a research engineer at OpenAI with eight years of experience at the intersection of machine learning, privacy, and systems. His PhD work at UCSD and internships at Meta, Tesla, and Tumult Labs center on quantifying model memorization and designing provable, application-specific privacy mechanisms for settings from LLM embeddings to location and social graph data. He has demonstrated practical attacks—such as diffusion-based reconstruction of training images for self-supervised vision models—and proposed mitigation and privacy accounting techniques that inform real-world deployment. Comfortable moving between theory and engineering, he has shipped algorithms, testbenches, and production-oriented designs from analog ADC blocks to ML privacy tools. Based in San Francisco, he blends rigorous academic publication with hands-on implementation at startups and major labs, often surfacing subtle privacy risks that are overlooked in standard evaluations.
8 years of coding experience
3 years of employment as a software developer
University of California, San Diego
B.S., Electrical Engineering & Signal Processing, B.S., Electrical Engineering & Signal Processing at Brown University
M.S., Computational Science and Engineering, M.S., Computational Science and Engineering at Harvard University