Summary
Bhishma Dedhia is a machine learning researcher-engineer with nine years of experience building agents and scalable inference systems, currently a Member of Technical Staff at Traversal designing deep research agents that diagnose and repair production-scale software. He holds a PhD from Princeton where his work advanced adaptive inference and scalable training for large generative diffusion and autoregressive models across language, vision, and time-series modalities. Bhishma’s background blends foundational theory—from restless bandits and RL lower bounds—to practical systems, including industry collaboration at Adobe on test-time scaling for video generation. Known for rigor and a taste for the “zen” of agent design, he focuses on making exabyte-scale systems navigable and self-healing. Based in New Jersey, he pairs top academic pedigree and valedictory honors from IIT Bombay with hands-on engineering to move cutting-edge ML into production.
9 years of coding experience
6 years of employment as a software developer
Indian Institute of Technology Bombay
Doctor of Philosophy - PhD, Machine Learning, Doctor of Philosophy - PhD, Machine Learning at Princeton University
Hindi, Gujarati, Marathi, English