Sanjit Bhat

Ph.D. student & software engineer

prof-pic.jpg

I’m a third-year CS Ph.D. student at MIT PDOS, advised by Frans Kaashoek and Nickolai Zeldovich, and supported by an NSF GRFP fellowship. I enjoy hacking on big computer systems, especially when that results in increased security, reliability, and correctness. Recently, I’ve been excited about the use of formal verification to guarantee such properties.

Prior to MIT, I was a Turing Scholar honors computer science undergraduate at UT Austin, advised by Hovav Shacham. Before that, I spent three years in the MIT PRIMES program, working with Srini Devadas and Aleksander Mądry.

Outside of academia, I interned at Amazon’s automated reasoning research group and at Gradient, a deep tech cybersecurity and identity infrastructure startup.

publications

  1. Formal verification of the Linux Kernel eBPF verifier range analysis
    Sanjit Bhat, and Hovav Shacham
    Usenix Symposium on Operating Systems Design and Implementation (OSDI) poster session., Jul 2022
  2. Var-CNN: a data-efficient website fingerprinting attack based on deep learning
    Sanjit Bhat, David Lu, Albert Kwon, and Srinivas Devadas
    Privacy Enhancing Technology Symposium (PETS)., Jun 2019
  3. Tech Report
    Towards efficient methods for training robust deep neural networks
    Sanjit Bhat, Dimitris Tsipras, and Aleksander Mądry
    Regeneron Science Talent Search (STS)., Feb 2019
  4. DynaFlow: an efficient website fingerprinting defense based on dynamically-adjusting flows
    David Lu, Sanjit Bhat, Albert Kwon, and Srinivas Devadas
    ACM CCS Workshop on Privacy in the Electronic Society (WPES)., Oct 2018