Vignesh Ungrapalli

Software Engineer

Bengaluru, Karnataka, India11 yrs 8 mos experience

Highly Stable

Key Highlights

Expert in video/image processing and optimization.
Proven experience with CUDA and deep learning algorithms.
Strong background in developing multimedia applications.

Stackforce AI infers this person is a SaaS expert specializing in multimedia processing and optimization.

Contact

Skills

Core Skills

Video/image ProcessingComputer VisionAudio Processing

Other Skills

CLinuxProgrammingPythonTensorflowVideo ProcessingCompression TechniquesOptimizationCUDAPerformance OptimizationObject TrackingOptical FlowDeep LearningDriver DevelopmentWindows

About

Software Engineer with a passion for everything in technology. I like hacking around with GPUs pushing them to the limits specifically in areas of Image/Video processing and Vision tasks.

Experience

Nvidia

Senior System Software Engineer

Jul 2014 – Present · 11 yrs 8 mos · Bengaluru, Karnataka, India

I work in the Multimedia team responsible for Video Drivers, Video SDKs and Video/Image/Audio processing applications.
Computer Vision & Applied Deep Learning:
Building bandwidth efficient Video Compression for Video Conferencing.
Worked on Scaling and Optimization of Audio Denoiser for Nvidia Maxine Cloud Video Conferencing Platform.
Worked on CUDA performance optimization for FRUC kernels.
Worked on Design and Implementation of high performance Object Tracker using Nvidia Optical Flow Engine.
Worked on optimization and implementation of libraries for Audio processing tasks such as Wake-Word and Text to Speech for robotics and nvidia drive.
Worked on Convolutional Autoencoder based Algorithms for Post processing Videos for better video quality and in real-time. Optimized (memory and run time) and Productized the algorithms as part of Nvidia NGX (Neural Graphics Experience) platform.
Video Drive Work:
Driver support for Miracast on Windows 10.
Worked on Video Decode SDK as part of Nvidia Video SDK (https://developer.nvidia.com/nvidia-video-codec-sdk). This would expose Nvidia hardware decode and postprocess capabilities to developers on Windows and Linux.
Worked on GPU Virtualization. This work was in conjunction with MSFT and other teams in NVIDIA which would help applications utilize GPU hardware capabilities in the container environment.