A researcher, scientist and implementor, Tom build systems that illustrate new ideas.
Tom did his undergraduate BSEE at the University of Virginia. He went to Carnegie Mellon University and completed a MS and PhD in Computer Engineering, under advisor Randy Bryant. His dissertation defined a paradigm for parallel computing called “Match and Move.” It is closely related to “Map and Reduce”.
At NASA Ames, Tom continued his work on parallel computing and compilers for supercomputers working with Rob Schreiber. In 1995 he joined Rambus to work on solving the increasing memory bandwidth gap microprocessors were facing, bringing experience with supercomputing algorithms to the memory systems of GPUs. His roles alternated between architecture and verification efforts.
Tom co-founded Sensr.net with Adam Beguelin and Yacin Bahi in 2009 to connect consumer surveillance cameras to the cloud, providing video analysis and archiving for the millions of hours of internet video these devices produce. This effort required the development of an innovative distributed platform for keeping thousands of TCP connections open with image traffic load balanced across computer vision processors. Sensr.net was a pioneer in the IOT where T=”Cameras”.
At SightCall, a WebRTC PAAS, Tom defined cloud APIs and built out the recording and transcoding capabilities of the platform.
Tom has produced numerous publications and patents. He enjoys speaking at conferences and blogs on a variety of subjects.
SightCall Blog Articles
Sensr.net Blog Articles