Deepak Singh Mohan
Machine Learning Engineer @Lime
Signup · Get unlimited contacts
WORK HISTORY
Machine Learning Engineer @Lime
A medical technology company specializing in real-time exercise applications for mobile devices• Continuous development of AI models for live feedback of therapy exercises• Developing and refining deep learning models for pose estimation on mobile devices to improve exercise tracking• Optimized ML training pipeline, increasing the training speed by 5x times• Proficiency in using CVAT application for bounding box, label and key point annotation videos on thecustom dataset• Setup a compliant system for systematically gathering training and test data in accordance with the EU AI Act• Logging, monitoring, and reporting experiments in ClearML to enhance experi•ment tracking and facilitate better reproducibility
EDUCATION
Visvesvaraya Technological University
Bachelor of Engineering (B.Eng.), Electronics and Communications Engineering
Technische Universität Darmstadt
Master of Science (M.S.), Information and Communication Engineering
ABOUT DEEPAK SINGH MOHAN
A passionate Data Scientist with over four years of experience specializing in training deep learning models for image processing tasks like whole-body segmentation, object classification, pose estimation and gesture detection. My expertise includes optimizing models like MobileNetV3 for real-time whole-body image segmentation on mobile devices and using heavyweight models like ResNet, HRNet for high-quality human segmentation. I possess strong proficiency in dataset management, encompassing in-depth analysis, meticulous preprocessing, and the creation of custom datasets. Additionally, I excel in adapting existing datasets like COCO, Openpose, Hagrid, and Youtube8M to meet specific project requirements through thoughtful customization.I am deeply dedicated to research, constantly exploring state-of-the-art algorithms, and developing novel training approaches, all while reducing model complexity through profiling and pruning to make the models more efficient and accessible. I am committed to pushing the boundaries of computer vision and contributing innovative solutions to the field.
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.