Jaydeep Borkar
(he/him/his)
PhD Student
Research interests
- Privacy and safety in language models
Education
- BS in Computer Engineering, Savitribai Phule Pune University — India
Biography
Jaydeep Borkar is a doctoral student in the Khoury College of Computer Sciences at Northeastern University, advised by David A. Smith.
Borkar's doctoral research, which he began in 2021 and expects to complete in 2026, focuses on safety and privacy in language models, such as memorization of training data. He is a founding organizer of the Trustworthy ML Initiative, which aims to lower the entry barriers into trustworthy machine learning.
Outside of academia, Borkar is a visiting researcher at Meta Superintelligence Labs in New York City, where he works on safety and alignment in large languade models.
Labs and groups
Recent publications
-
Privacy Ripple Effects from Adding or Removing Personal Information in Language Model Training
Citation: Jaydeep Borkar, Matthew Jagielski, Katherine Lee, Niloofar Mireshghallah, David A. Smith, Christopher A. Choquette-Choo. (2025). Privacy Ripple Effects from Adding or Removing Personal Information in Language Model Training CoRR, abs/2502.15680. https://doi.org/10.48550/arXiv.2502.15680 -
Mind the Gap: Analyzing Lacunae with Transformer-Based Transcription
Citation: Jaydeep Borkar, David A. Smith. (2024). Mind the Gap: Analyzing Lacunae with Transformer-Based Transcription CoRR, abs/2407.00250. https://doi.org/10.48550/arXiv.2407.00250 -
Simple Transparent Adversarial Examples
Citation: Jaydeep Borkar, Pin-Yu Chen: Simple Transparent Adversarial Examples. ICLR 2021 Workshop on Security and Safety in Machine Learning Systems.