Embedded Files
About Me
I’m an AI researcher and engineer with a PhD in Imaging Science and ~8 years of experience in artificial intelligence, computer vision, and deep learning. My work spans both research and industrial settings, with a focus on building efficient, scalable, and deployable AI solutions.
Most recently, I worked as an R&D Engineer at Digimarc, where I optimized digital watermarking pipelines using a combination of image processing, computer vision, and deep learning. Prior to that, during my PhD at RIT and internships at Mitsubishi Electric Research Laboratories and Applied Image, I focused on multimodal sensor fusion (RGB, IR, hyperspectral, multispectral, LiDAR, SAR) and model compression techniques (low-rank factorization, pruning, quantization) to enable lightweight neural architectures for deployment in resource-constrained environments and edge devices.
I’m passionate about advancing AI systems that are not just powerful but also efficient, robust, and adaptable to real-world challenges.
Experience
Experience
Research and Development Engineer, Digimarc Corporation, OR, US, January 2024 - May 2025.
Research Assistant (during Ph.D.), Rochester Institute of Technology, NY, US, August 2019 - January 2024.
Research Intern, Mitsubishi Electric Research Laboratories, MA, US, August 2022 - December 2022.
Computer Vision and Deep Learning Intern, Applied Image Inc., NY, US, May 2020 - August 2020.
Teaching Assistant (during Ph.D.), Rochester Institute of Technology, NY, US, August 2018 - July 2019.
Data Scientist, Climate Connect Technologies, Pune, India, August 2016 - July 2018.
Education
Education
Ph.D. Imaging Science, Rochester Institute of Technology (RIT), August 2018 - January 2024.
M.S. Physics, Indian Institute of Technology (IIT), Delhi, India, July 2014 - June 2016.
B.Sc. Physics, University of Delhi, India, July 2011 - June 2014.
Research Interests
Adaptive rank determination for lightweight tensorized neural networks for mobile/edge applications.
Cross-modal prior knowledge sharing strategy to reduce hypothesis space for multimodal domain adaptation.
Progressive/incremental representation learning on streaming data for image reconstruction and denoising.
Tensor factorization/modelling of convolutional kernels for network compression, optimization, and speed-up.
Limited multi-modal data fusion of unpaired unimodal legacy networks for classification.
Optimized CNN for detection of small oriented objects for remote sensing applications.
Multi-modal data fusion for aerial perception (classification, detection, and segmentation).
Page updated
Google Sites
Report abuse