Advisory Board Member for CS Department at University of Colorado

I led a large R&D team of researchers and engineers at NEON (an independent business venture within Samsung Electronics) pioneering and developing Artificial Human technology for wide ranging applications. We developed novel AI enabled products and services for various business verticals.For more details about NEON, visit https://neonlife.ai/ or follow @neondotlife on any social platform

At Samsung Electronics, I was leading a large team of computer vision and machine learning researchers responsible for building AI/AR Camera platform built into Samsung phones and used by millions of users daily. The work started with Bixby Vision and matured into broader AI/AR camera platform for Samsung Galaxy phones

I am member of Think Tank Team responsible for various computer vision related activities at Samsung Research America in Mountain View, CA. (http://thinktankteam.info)

I am member of Think Tank Team at Samsung Research America in Mountain View, CA. (http://thinktankteam.info)

Operational Issues in Large Scale Object Recognition: More specifically I am developing methods for systems that allow scalable and dynamic addition of new object categories at training time and are robust to unseen categories at testing time. I have also developed methods for face recognition in streaming settings and handling missing data at run time.

Developed wide range of algorithms for detecting actionable content in Social. Responsible for developing online and offline classification system, active learning systems, detecting trending content etc on Twitter Data. Led research efforts in machine learning for a team of 6 people (4 PhDs).

Developed a novel personal photo-collection organization system. Wrote algorithms to summarize large personal photo-collections into context specific picture storylines.

Developed methods for incremental learning for face recognition system. Our system allows fast updates for pre-trained system. Such systems are crucial when minimum downtime is desired.

Developed methods for incremental learning for face recognition system. Our system allows fast updates for pre-trained system. Such systems are crucial when minimum downtime is desired.

Developed supervised classification algorithms for analyzing Social Media Data (NLP). Developed feature extraction and learning infrastructure (online learning, SVMs, large-scalemulti-class classification, batch learning). Developed framework for large scale online/incremental learning.Proposed solutions, developed code that went in production and developed machine learning based product strategy. Also worked on conversation analysis in Social.

Developed tools and techniques to take computer vision to masses. Developed a novel platform called VisionBlocks (published in SocialCom 2011). VisionBlocks also received Google's App Engine Research Award, awarded for most innovative software projects from all over the world. Developed Business plan for MIT 100K Business competition for my work. Organized demos and workshops for Cambridge Science Festival for MIT Media Lab as a part of science outreach.

Worked on wide ranging problems in computational neuroscience with emphasis on deep neural networks and supervised/unsupervised feature learning. Special emphasis was given on establishing benchmarks for object recognition and pedestrian detection systems as part of DARPA's Neovision effort. Developed and ported deep learning methods to GPUs, long before it was trendy. This work was done in McGovern Institute of Brain Research and MIT's Brain and Cognitive Science department with Dr. Nicolas Pinto and Prof. James DiCarlo (Head)

Developed first of its kind multi-fingerprint recognition system. My role included core algorithm design and research, software development and evaluation. Closely collaboratedwith Securics Inc., a local company specializing in privacy enhanced biometrics. Also developed system for predicting failure in recognition systems.