Developed Monocular Depth Estimation and 3D Image Reconstruction models for VR/AR

Leveraged semantic segmentation masks to refine depth boundaries for high-fidelity 3D reconstruction on mobile chipsets

"Weights visualization of 3D ConvNets for 3D data"

Visualize the weights (or filters) of CNN trained on 3D data. Simple three-layer model architecture is used for training purpose and the visualization approach is based on the gradient ascent optimization technique. An effort is made to see the influence of desired CNN filter on a random noisy image by searching for the image pixels that maximize the output of the filter.

HEDGE as L1, TDD L1 TX

▪ Expertise: C Programming, Algorithms & Data Structures, Test Suite Development, Microsoft Office suite.

▪ Major Work Done: Developed a unit-testing tool to verify the working of inter- layer messaging covering all scenarios in modem protocol stack.

▪ Other Achievements: Appointed as the office floor representative to management for 400 engineers. Main responsibility along with scientific work was to organize meetings with HR managers and arrange monthly outings for teams.

Wireless Modem Technology software development in 3G Layer1(Physical layer).

▪ Majorly worked on scripting test cases with Perl and C to verify the functionality of 3G Physical layer procedures and troubleshooting issues related to call connectivity and data throughput.