Jagadeesh Sankaran
Senior Director, Computer Vision Architecture, PVA at NVIDIA
Dublin, California, United States
Summary
🤩
Rockstar🎓
Top SchoolJagadeesh Sankaran is a seasoned computer vision and video architecture leader with over two decades of experience designing high-performance vision accelerators and codecs, and seven years in senior roles driving NVIDIA's PVA strategy for Orin and Xavier. He combines deep hardware-software co-design expertise from his TI work on the Embedded Vector Engine and HD video engines with modern ML deployment skills, having contributed a YOLO plugin and TensorRT/DeepStream integrations used in NVIDIA AI IoT reference apps. Based in Dublin, CA, he blends research rigor (PhD in EEC) with pragmatic product delivery, repeatedly moving complex vision algorithms into efficient silicon and software stacks. An IEEE Senior Member and proven architect, he is particularly adept at creating specialized data paths and post-processing pipelines that squeeze maximum throughput from constrained accelerators.
7 years of coding experience
13 years of employment as a software developer
Doctor of Philosophy (Ph.D.), Electrical, Electronics and Communications Engineering, Doctor of Philosophy (Ph.D.), Electrical, Electronics and Communications Engineering at The University of Texas at DallasBachelor of Technology (B.Tech.), Electronics and Communication, 1st class, Bachelor of Technology (B.Tech.), Electronics and Communication, 1st class at National Institute of Technology WarangalMasters Degree, Electrical Engineering, Masters Degree, Electrical Engineering at Louisiana State University
Github Skills (10)
tensorrt10
mle10
object-detection10
nvidia10
computer-vision10
c-language10
cprogramming-language10
ml10
gstreamer9
cuda8
Programming languages (1)
Github contributions (1)
NVIDIA-AI-IOT/deepstream_reference_apps
Jul 2018 - Feb 2019
Samples for TensorRT/Deepstream for Tesla & Jetson
Role in this project:
ML Engineer
ML EngineerContributions:48 commits, 24 pushes, 1 tag in 7 months
Contributions summary:Jagadeesh primarily contributed to the development and integration of a YOLO (You Only Look Once) object detection plugin for the DeepStream framework. Their work involved implementing the YOLO plugin, including creating the necessary code for model integration, bounding box parsing, and metadata handling. The user also created a standalone TensorRT application, which can run YOLO models, and improved the overall efficiency of the post-processing steps. Further work involved refactoring the project structure, fixing bugs and updating the code to be consistent with DeepStream 3.
jetsondeep-learningdeepstreamtensorflowtensorrt
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.