Current: foundational research unitPast: computer vision @ google research, camera algorithms @ pixel

Working on CV algorithms to enable a multiplayer AR experience on phones.

Building a point cloud annotation tool to semi-autonomously annotate point clouds using a fusion of geometric and learned features.More information here: https://mscvprojects.ri.cmu.edu/2018team10/

Research internship on deep learning based image stylisation.* Built a deep learning system to perform a style transfer where the style can be parameterised* Wrote custom tensorflow layers that implement the renderer function for the parameterised style. * Invented novel methods to enable embedding of non differentiable renderer functions in a deep network*Built a new dataset by analysing 10k+ movie trailers for extracting datapoints suitable for trainingMentors: Kevin Wampler, Oliver Wang, Kalyan Sunkavalli, Eli Shechtman

Worked with the PWC Risk & Regulatory Services Innovation Center. I worked on the Resilience Analysis and Design on IoT-Based Smart Infrastructures project. My work primarily involved C++ development for moulding the theoretical work done around IOT risk assessment into a tool to be used on the field for evaluating the vulnerability and preparedness of large-scale clients against IOT attacks.Project PI: Prof. Bruno Sinopoli

I work in the eLearning group at Adobe Systems. Previously, I worked on the background replacement feature in Adobe Presenter Video Express, where we developed a novel algorithm to replace backgrounds without the need of a green screen. Work involved devising new methods to segment the video, achieving pixel perfect matting and providing a visual feedback to the user pre-recording on the expected quality of the output. I am currently experimenting with providing eLearning content in Virtual Reality.

Worked in the Advanced Software Division at the Center of Excellence in Bangalore.During the course of my internship, I was involved in the following three projects:Project I:Conceptualised and developed a solution for monitoring cloud infrastructure alongside on-premise infrastructure, to provide a single pane of glass view of the resources to the administrator. Worked with the monitoring SDKs for AWS and Google Cloud and developed a Solution Pack for the EMC ViPR-SRM.Project II: Delivered on an internal tool to be used by developement teams across the globe. Work involved writing parsing and validation logic for the tool which read and set elements from XML files used in the product ViPR-SRM. Also worked with the Swing library in Java to develop the UI. Project III:Worked on development tasks towards the release of ViPR SRM. My worked was with the Solution pack for supporting the cloud storage system - EMC Atmos in ViPR SRM. This project helped develop a good understanding of the Data storage domain.

Worked on a research project which involved coming up with solutions for the open-ended problem of background extraction and substitution in Videos. During the first half of the internship we spent a substantial time towards a thorough research on the basic and advanced concepts in Computer Vision with an emphasis on Background Replacement. The second half involved developing a novel solution to replace video background with minimal interaction from the user. We developed a graph-cut based algorithm which used an ensemble of classifiers trained on different cues (colour, motion, shape and human features) working in tandem with each other to give the final output.

Learning Centre, Operations.• Ran daily class and remedial sessions for a batch of 25 high school students• Liaised with school partners to strengthen Avanti’s relationship • Helped create Avanti's outreach strategy for the year 2013-2014