Graduate Research Assistant
Pennsylvania State University | January 2020 - Present
Goal-Oriented Information Extraction in Spoken Dialogue Systems ∠ Developed a novel method1 to extract frame-semantics related to information disclosure in conversations system. ∠ Trained semi-supervised adversarial model ensembles for Reddit text classification.
Samsung | June 2015 - July 2019
Lead Engineer | Healthcare AI (Research) | Vision Intelligence Team Apr 2018 - Jul 2019 ∠ Developed remote health monitoring algorithms for Samsung Robot using Computer Vision, Signal Processing. ∠ Visual PPG : Developed a novel algorithm for heart signal estimation using temporal tracking of skin-clusters in face video. ∠ Implemented classification models for arrhythmia detection using Visual PPG data collected from 250 subjects. #### Project: Fashion Recognition | Vision Intelligence Team Jun 2017 - Mar 2018 ∠ Developed a live fashion recommendation system for E-commerce application in Samsung Smart TV. ∠ Trained Faster R-CNN with contrastive loss for recognizing clothing of 11 classes with a 45% recall @ top-5 retrieval. ∠ Implemented a “Street Image to In-Shop” clothing retrieval deep learning framework. ∠ Integrated the clothing retrieval framework to Samsung TV Content Recognition system deployed on AWS. #### Project: Automated Content Recognition (ACR) | Data Intelligence Team Jul 2015 - May 2017 ∠ Developed a dashboard monitoring system for ACR Infra using Node.js and NGINX API. ∠ Optimized the Kafka-Cassandra pipeline and ElaticSearch query templates to reduce latency by 25x. ∠ Obtained a reduction of over 40% in the total ACR network traffic by implementing Redis caching module. ∠ Awarded Best Employee of the Month for facilitating rapid deployment using AWS CloudFormation in US region.
Pennsylvania State University
Computer Science & Engineering, MS | August 2019 - Present
Computer Science and Engineering, B.Tech | June 2011 - May 2015