Bhishma Dedhia

CTO

Princeton, New Jersey, United States6 yrs 4 mos experience

Highly Stable

Key Highlights

PhD from Princeton focused on generative models.
Expertise in scaling machine learning for real-world applications.
Research bridging theory and practical utility in diverse fields.

Stackforce AI infers this person is a Machine Learning Researcher with expertise in scalable models for Healthcare and Public Policy.

Contact

Skills

Core Skills

Software EngineeringResearch And DevelopmentMachine LearningResearch

Other Skills

deep research agentsproduction-scale softwareadaptive inference methodsscalable training objectiveslarge-scale generative modelsmultimodal modelsTest-Time InferenceVideo Generation Modelsrestless multi-armed banditsrandomized algorithmsresource allocation problems

About

Building and understanding the science (and zen) behind agents that can navigate exabyte-scale software systems. Previously, I received a PhD from Princeton, where I primarily focused on scaling generative models and improving their test-time capabilities. My work has bridged theory with real-world utility across healthcare, public policy, and designing RCTs. https://bhishmadedhia.com/

Experience

6 yrs 4 mos

Total Experience

2 yrs 10 mos

Average Tenure

7 mos

Current Experience

Traversal

Member of Technical Staff

Sep 2025 – Present · 7 mos · New York City Metropolitan Area · On-site

Designing deep research agents capable of understanding and repairing production-scale software.

deep research agentsproduction-scale softwareSoftware EngineeringResearch and Development

Adobe

Research Scientist Intern

May 2024 – Nov 2024 · 6 mos · San Francisco Bay Area · On-site

Scaling Test-Time Inference for Video Generation Models, https://arxiv.org/abs/2503.17539

Test-Time InferenceVideo Generation ModelsMachine Learning

Princeton university

Graduate Researcher

Sep 2020 – Aug 2025 · 4 yrs 11 mos · Princeton, New Jersey, United States

Research on adaptive inference methods and scalable training objectives for large-scale generative diffusion and autoregressive models. My work bridged foundational machine learning with strong empirical performance to scale inference, enhance capabilities, and improve the efficiency of multimodal models across language, vision, and time-series data.
https://arxiv.org/abs/2503.17539; https://arxiv.org/abs/2403.07887; https://arxiv.org/abs/2305.17328
https://arxiv.org/abs/2305.17262; https://arxiv.org/abs/2207.04208; https://arxiv.org/abs/2205.11656, https://arxiv.org/abs/2507.13966

adaptive inference methodsscalable training objectiveslarge-scale generative modelsmultimodal modelsMachine LearningResearch

Indian institute of technology, bombay

Undergraduate Researcher

Aug 2019 – Jun 2020 · 10 mos

Research on restless multi-armed bandits and randomized algorithms to solve large-scale resource allocation problems in multi-sensor systems, working with the Stochastic Systems Lab. Also dabbled in proving cool theoretical lower bounds for reinforcement learning.
https://arxiv.org/abs/2005.04036v2; https://arxiv.org/abs/1911.12842; https://arxiv.org/abs/2009.07842

restless multi-armed banditsrandomized algorithmsresource allocation problemsResearch