Jishnu Roychoudhury

CS+Math @ Princeton University

Princeton, New Jersey, United States

About

Passionate about Artificial Intelligence, excited to explore both their capabilities and opportunities for alignment of these systems.

Experience

  • Incoming Algorithm Developer (Quantitative Researcher) Intern at Hudson River Trading
    Oct 2025 - Present · 10 mos

  • Princeton University (On-site)
    • Research Assistant
      Nov 2023 - Present · 2 yrs 9 mos

      • Designing and implementing first fast GPU kernel for Attention Jacobian Vector Product with Prof. Tri Dao (Jun '25 – Aug '25) • Developed and analyzing bidirectional shortest-path algorithms and contextually efficient heap data structures with Prof. Robert Tarjan (Turing Award '86) (Jun '25 – Aug '25) • Designed and implemented "disentangled" transformer formulation which restricts attention to isolate position/content sub-circuits (Nov ’23 – May ’24)

    • Teaching Assistant
      Aug 2023 - May 2025 · 1 yr 10 mos

      • COS 423 (Theory of Algorithms) (Spring '25) • COS 597C (Graduate Natural Algorithms) (Fall '24) - Head TA • COS 226 (Data Structures and Algorithms) (Fall '23, Spring '24)

  • Software Engineer at Hillspire
    May 2024 - Aug 2024 · 4 mos

    Worked on video generation models and applications with Eric Schmidt (former CEO of Google) and Sebastian Thrun (founder of Waymo, GoogleX, Udacity). • Built high-throughput video generation backend on AWS, optimized for low-latency media delivery. • Engineered a parallelized pipeline on Lambda for video synthesis, achieving a 32× speedup over the baseline architecture. • Developed embedding-based recommender with Bayesian collaborative filtering, boosting average engagement time by 65%.

  • Capability Development Group at Singapore Armed Forces (SAF)
    Aug 2021 - Aug 2023 · 2 yrs 1 mo

    • Led hardware-software IT security project in partnership with IBM and Singapore DSTA. • Invented device communication protocol based on dynamic programming state compression, enabling continued use of legacy technology platform, driving cost savings. • Co-ordinated varied stakeholders and project members, improving my communication and high-level design skills. Overall Performance Rating: Outstanding (max) Overall Conduct Rating: Outstanding (max)

  • Research Intern at Singapore University of Technology and Design (SUTD)
    Jun 2021 - Dec 2021 · 7 mos

    Research in applications of computer vision to security.