John Abascal

Computer Science PhD Student at Northeastern University

me.png

177 Huntington

Boston, MA 02115

I am a fourth-year Computer Science PhD Student at Northeastern University, where I am advised by Jonathan Ullman and Alina Oprea. I completed my B.S. in Pure Mathematics at Florida State University and conducted my undergraduate honors thesis under the guidance of Alec Kercheval and Nathan Crock.

My primary research interest is privacy in deep learning. More broadly, I am interested in both theoretical computer science and applied machine learning, including topics such as differential privacy, adversarial machine learning, and the trustworthiness of machine learning algorithms.

I have also held industry positions in both software engineering and data science at Intuit and NewSci Labs, respectively. During the summers of 2023 and 2024, I was an Applied Research Intern on LinkedInโ€™s Data Privacy team, where I worked on empirically measuring privacy leakage in analytics and LLMs. This semester, I am a Student Researcher at Google, exploring open problems in machine unlearning ๐Ÿ”’

news

Jun 1, 2024 New paper on poisoning retrieval-augmented generation (RAG) systems! ๐Ÿค–
Feb 1, 2024 Our paper was accepted at PETS 2024! ๐ŸŽ‰
Jun 20, 2023 Our work on privacy leakage in finetuned models was accepted into the 2023 ICML AdvML Frontiers Workshop! ๐ŸŽŠ
Jun 4, 2023 New paper on privacy leakage in finetuned vision and language models!
Dec 6, 2022 Our work on efficient property inference attacks has been accepted into IEEE S&P 2023! ๐ŸŽ‰