Devendra Jangid
Senior Research Engineer
Samsung Research America
Email Id: dkjangid@ucsb.edu
Resume Projects Github Google Scholar
Introduction
I am a Senior Research Engineer at Samsung Research America (SRA). I have done Ph.D. in Electrical and Computer Engineering from University of California, Santa Barbara, under the supervision of Prof B.S. Manjunath. I got my bachelor's degree in Electrical Engineering from Indian Institute of Technology (IIT), Jodhpur in 2015. I have worked as a Scientist B from 2015-2018 in Research Center Imarat, Defense Research and Development Organization (DRDO).
Research Interests:
Machine Learning, Deep Learning, Generative Models
Image Processing and Computer Vision
AI for Science
Education
University of California, Santa Barbara (June, 2020- March, 2024 )
Ph.D. in Electrical and Computer Engineering
Advisor: Prof B.S. ManjunathUniversity of California, Santa Barbara (Sept, 2018- June, 2020)
M.S. in Electrical and Computer EngineeringIndian Institute of Technology, Jodhpur (July, 2011- June, 2015)
B.Tech. in Electrical Engineering
Work Experience
Samsung Research America, Plano, TX (April'24 - Current)
Intelligent Imaging, Mobile Processor Innovation (MPI) LabWorking on Gen-AI technology for Samsung Smartphone Camera
Defense Research and Development Organization (2015-2018)
Scientist B, Research Center Imarat Lab
Object Detection and Tracking Algorithm
Implemented image processing algorithms on embedded hardware systems for object detection and tracking on infrared images.
Internships
Samsung Research America, Plano, TX (Summer'22)
Research Intern, Computational Camera Team, Mobile Processor Innovation (MPI) Lab
Mentors: Dr. Seok-Jun Lee, Dr. Hamid R. Sheikh
Project: Efficient Single Image Super-resolution for Samsung Smartphone Cameras (Patent in progress)
Designed an efficient single image super-resolution deep learning framework for Samsung smartphone camera.
Samsung Research America, Plano, TX (Summer'21)
Research Intern, Computational Camera Team, Mobile Processor Innovation (MPI) Lab
Mentors: Dr. Seok-Jun Lee, Dr. Hamid Sheikh
Project: AI-Based Image Sharpening (Patent in progress)Designed an efficient deep learning framework to improve the night photography of Samsung smartphone camera.
Jacobs University, Bremen, Germany (Summer'14)
Research Intern (DAAD-WISE Scholar), Transmission Systems Lab
Mentors: Prof. Dr. -Ing. Werner Henkel, Dr. Khodr Saaifan
Project: Impulse Noise Detection in Cooperative CommunicationWorked on 5G Cooperative Communication. Implemented ML/MAP estimation algorithms to find the effects of impulse on wireless received signals.
Publications/Patents
Devendra K. Jangid, Neal R. Brodnik, Michael G. Goebel, Amil Khan, SaiSidharth Majeti, McLean P. Echlin, Samantha H. Daly, Tresa M. Pollock, B.S. Manjunath, "Adaptable Physics-Based Super-Resolution for Electron Backscatter Diffraction Maps", Nature Publication Journal, Computational Materials 2022
[Paper] [Code] [Bisque Module] [Poster]
Devendra K. Jangid, Neal R. Brodnik, McLean P. Echlin, Tresa M. Pollock, Samantha H. Daly, B.S. Manjunath "Q-RBSA: High-Resolution 3D EBSD Map Generation Using An Efficient Quaternion Transformer Network",Nature Publication Journal, Computational Materials 2024
[Paper] [Code] [Bisque Module]
Devendra K. Jangid, Neal R. Brodnik, McLean P. Echlin, Samantha H. Daly, Tresa M. Pollock, B.S. Manjunath, "Titanium 3D Microstructure for Physics-based Generative Models: A Dataset and Primer", SynS and ML Workshop, International Conference on Machine Learning , 2023 (Best Paper Award and Spotlight Presentation)
[Paper] [Poster] [Presentation]
Devendra K. Jangid, Neal R. Brodnik, Amil Khan, Michael G. Goebel, McLean P. Echlin, Tresa M. Pollock, Samantha H. Daly, B.S. Manjunath "3D Grain Shape Generation in Polycrystals Using Generative Adversarial Networks", Integrating Materials and Manufacturing Innovation 2022
[Paper]
Devendra K. Jangid, Neal R. Brodnik, Amil Khan, Michael G. Goebel, McLean P. Echlin, Tresa M. Pollock, Samantha H. Daly, B.S. Manjunath, "3DMaterialGAN: Learning 3D Shape Representation from Latent Space for Materials Science Applications", arXiv 2020
[Paper]
Devendra K. Jangid, B.S. Manjunath, "Enhancing Materials Microstructure Analysis with Physics-Informed Computer Vision", 2024
[Thesis]
Devendra K. Jangid, Seok-Jun Lee, Hamid R. Sheikh, "Machine learning-based approaches for synthetic training data generation and image sharpening, US Patent 2022
[Patent]
Devendra K. Jangid, Abhiram Gnanasambandam, John, Seok-Jun Lee, Hamid R. Sheikh, "Efficient Single Image Super-resolution for Smartphone Camera", US Patent 2022 (Under Review)
Projects
Please Check Projects Page for Details
Honors and Awards
Received DAAD-WISE (German Academic Exchange Service) Scholarship from the German Federal government for a summer internship in 2014
Appointed to SCIENTIST post by President of India to work in Defense Research and Development Organization (DRDO), India in 2015
Best Paper Award in Synergy of Scientific Modeling and Machine Learning Workshop at ICML 2023
Outstanding Teaching Assistant award from ECE department of UC, Santa Barbara in 2019
Received Fellowship from ECE department, UC, Santa Barbara for Fall 2018
Received ECE Dissertation Fellowship for Summer 2023
Received Doctoral Travel Grant from UCSB to attend ICML'23 Conference
Merit cum Means (MCM) scholarship for 4 years (2011-2015) in undergraduate at IIT, Jodhpur
Technical Skills
Programming Language: Proficient in Python, Experience in C++, C, and Bash
Packages: Proficient in PyTorch, NumPy, Sci-Kit, and Latex, Experience in TensorFlow
Platform: Proficient in Linux and Windows, Experience in Docker
Version Control System: Proficient in Git
Teaching Assistant
Conducted lectures and tutorials:
Introduction to Deep Learning, Digital Communication, Introduction to ECE, Probability and Statistics
Relevant Course Works
Introduction to Deep Learning (A+), Advanced Computer Vision, Digital Image Processing, Pattern Recognition, Theoretical Machine Learning, Optimal Estimation and Filtering (A+) , Matrix Analysis and Computation, Information Theory