Usha

Developed and deployed real-time computer vision systems for industrial environments, focusing on video annotation workflows, dataset preparation, model training, and optimized inference pipelines. Worked on safety monitoring and asset inspection use cases, including action recognition and unsafe behavior detection, as well as asset quality inspection using deep learning-based object detection. Hands-on with YOLO-based object detection and Mask R-CNN instance segmentation using Python with PyTorch and OpenCV. Collaborated with teams to enable AI-driven vision deployments across edge and cloud.

Developed and deployed real-time computer vision systems for industrial environments, specializing in video annotation, model training, and optimized inference pipelines. Worked on action recognition and unsafe behavior detection, and built asset quality inspection using deep learning-based object detection. Produced and prepared datasets through extensive video annotation workflows, and implemented YOLO-based object detection and Mask R-CNN instance segmentation. Applied Python-based computer vision development using PyTorch and OpenCV to deliver scalable solutions for safety monitoring, asset inspection, and operational analytics.

Supported AI-driven vision deployments across edge and cloud environments by managing server bring-up, infrastructure configuration, and deployment of computer vision workloads on on-prem and cloud platforms. Performed Linux administration (installation, configuration, patching, monitoring) and network configuration. Deployed and configured cloud infrastructure for AI workloads on AWS and Azure, and assisted with troubleshooting deployment and performance bottlenecks. Helped install and maintain environments required for model deployment and operational readiness.