Amazon LAB126
Sunnyvale, CA
Senior Applied Scientist & Tech Lead
March 2015 - present - I lead research and development for new computer vision and machine learning features for Amazon products.
Mitsubishi Electric Research Lab (MERL) Cambridge, MA Research Intern Host: Fatih Porikli Summer 2013 - Proposed an algorithm for joint denoising and super-resolution of natural images.
- Proposed a layered model for texture preserving super-resolution of natural images.
Siemens Corporate ResearchPrinceton, NJ Research Intern Host: Ying Zhu Summer 2011 - Proposed a variational optimization framework for multimodal image registration using quadratic mutual information.
- Proposed a method for detection and localization of anatomical structures in medical images using regression forests.
UtopiaCompression Corporation Los Angeles, CA Research Intern Host: Jacob Yadegar and Anurag Ganguli Summer 2009 - Proposed an algorithm for biomedical signal modeling for anomaly detection using hidden Markov models.
- Developed adaptive triangulation/meshing algorithms for image compression and 2D surface crack modeling.
University of Illinois at Urbana-Champaign Urbana, IL Fellow, Research Assistant and PhD Candidate Advisor: Narendra Ahuja Aug 2010 - Feb 2015
- Developing transform domain algorithms for learning-based super-resolution of natural images.
- Developing algorithms for robust registration and dense correspondence estimation between image sequences.
University of Florida Gainesville, FL Research Assistant and MS Candidate Advisor: Jose Principe Fall 2008 - Spring 2010 - Proposed a family of robust cost functions for supervised learning (classification, regression, adaptive filtering) based on a novel similarity metric.
- Proposed information theoretic descriptors as cost functions for machine learning and signal processing algorithms.
Dhirubhai Ambani Institute of Information and Communication Technology Gandhinagar, India Undergraduate Student Advisor: Suman Mitra and Manjunath Joshi Fall 2004 - Spring 2008 - Proposed a sampling-resampling based Bayesian learning algorithm for learning mixture models for robust background subtraction in videos and image segmentation.
|
|
