About

DreamVu is an applied research company.

Founded January 2026. Building cameras since 2015. DreamVu captures real-world human data and delivers it as training data for physical AI, using a camera we designed, built, and own.

DreamVu began in 2015 inside the Centre for Visual Information Technology at IIIT Hyderabad. Prof. Anoop Namboodiri, his graduate student Rajat Aggarwal, and a fellow student set out to solve a problem the imaging industry considered impractical: capturing full 360° stereoscopic depth in a single shot, from a single sensor, with no seams or blind spots. Their answer was a new computational optics design, published at CVPR 2016 as Panoramic Stereo Videos with a Single Camera.

The company was founded later that year. Rajat and Anoop stayed on to build the technology. In 2018, IIIT Hyderabad, which owned the foundational IP, granted DreamVu an exclusive license in exchange for equity.

What followed was nearly a decade of camera engineering. PAL, our first-generation 360° stereo sensor, went from lab prototype in 2016 to production in 2019 and is covered by around 16 patents across optics, electronics, software, and calibration. Alia, our omnidirectional stereo camera, was prototyped in 2017, entered production in 2020, and is protected by around 14 more. DreamVu designed, built, and owns it.

In January 2026, Sashi Reddi joined as Chairman, and the four of us refounded the company as DreamVu.ai. The camera stopped being the product and became the instrument. We now use it to capture the real-world data that robot foundation models and world models train on.

We do not sell a catalog and we do not sell cameras. You bring us a data problem, we build a capture program around it, and we deliver in the format your training stack uses.

Palo Alto

California · US headquarters

Philadelphia

Pennsylvania · East coast office

Hyderabad

India · R&D center

Founding team

Sashi Reddi

Chairman & CEO

DreamVu is Sashi’s sixth company as CEO and co-founder. He founded AppLabs, which grew into the largest independent software testing company in the world with more than 2,500 staff across the US, UK, and India, backed by WestBridge and Sequoia India and acquired by CSC. His earlier companies were EZPower Systems, later part of Oracle, and FXLabs, later part of Electronic Arts. He has spent more than two decades building and investing in technology companies across the US and India. BTech from IIT Delhi and MS from NYU, both in computer science. PhD in operations and information management from Wharton.

Rajat Aggarwal

Chief Technology Officer

Rajat holds a BTech and Masters in computer science from IIIT Hyderabad, specializing in computer vision and computational photography. His research produced the first panoramic stereo camera built on a single sensor, which became the basis for DreamVu. He previously built medical optical devices with MIT Media Lab and the L V Prasad Eye Institute. He has over a decade of experience in camera hardware and software, with deep knowledge of optics, manufacturing, and production scaling, and holds multiple patents.

Scholar ↗

Prof. Anoop Namboodiri

Chief Science Officer

Anoop is Professor and Dean of Faculty Affairs at IIIT Hyderabad, working in pattern recognition, biometrics, and computer vision. He holds a PhD from Michigan State University. He designed the biometric deduplication system for India’s Aadhaar program, the largest biometric identity system in the world, covering more than a billion people. At DreamVu he leads work on omnidirectional 3D vision and the software behind it. He has published more than 75 peer-reviewed papers.

Scholar ↗

Dr. Parikshit Sakurikar

VP Imaging & AI

Parikshit holds a PhD in computer science from IIIT Hyderabad, where his research covered focus, defocus, and depth of field in wide-aperture imaging. He has more than 20 publications across computer vision and graphics, spanning depth estimation, light field synthesis, scene refocusing, and neural rendering. At DreamVu he runs imaging and AI, from camera optics through to the data pipelines and model fine-tuning.

Scholar ↗

Scientific advisors

Prof. Takeo Kanade

U.A. and Helen Whitaker University Professor of Robotics and Computer Science, Carnegie Mellon University

A foundational figure in computer vision and robotics. Takeo was founding director of CMU’s Robotics Institute from 1992 to 2001 and has written more than 400 technical papers and holds over 20 patents across stereo vision, structure-from-motion, and autonomous systems. His honors include the Kyoto Prize, the Benjamin Franklin Medal, and election to the US National Academy of Engineering. He has advised DreamVu since its early years.

Scholar ↗

Prof. Michael S. Brown

Professor and Canada Research Chair in Computer Vision, York University; VP, Samsung AI Center Toronto

Michael is a leading authority on the camera imaging pipeline, color science, and computational photography. His research has shipped in hundreds of millions of consumer devices, including the Samsung Galaxy line. He has served as program or general chair for CVPR, ACCV, 3DV, and WACV.

Scholar ↗

Prof. P.J. Narayanan

Professor, IIIT Hyderabad; Director Emeritus, 2013–2025

P.J. founded the Centre for Visual Information Technology at IIIT Hyderabad, now the largest computer vision and graphics research group in India and the lab where DreamVu’s technology began. Earlier, at CMU’s Robotics Institute, he built the Virtualized Reality system for capturing dynamic 3D scenes. His research covers 3D vision, computational cameras, and GPU computing.

Scholar ↗

Prof. Dinesh Manocha

Paul Chrisman Iribe Chair in CS & ECE and Distinguished University Professor, University of Maryland

Dinesh leads one of the most productive research groups in robotics, computer graphics, and AI for autonomous systems, with more than 850 publications and 19 patents. He co-founded Impulsonic, acquired by Valve, and is a Fellow of ACM, AAAI, AAAS, IEEE, and NAI. His group’s software is licensed to more than 60 commercial vendors.

Scholar ↗

Backed by

SRI Capital

Ben Franklin Technology Partners

Broad Street Angels

Tell us what your model needs to learn.

Capture programs, research collaboration, and dataset partnerships.

Talk to us