Harish Siva Subramanian

Data Scientist @Verra Mobility

Built Flask web services to automate the necessary image processing tasks using Deep learning, Computer Vision, and State of the art architectures like YOLO(on GPU) to create services like Vehicle/Plate, OCR detection, etc. • Collaborated with a team of SDEs, DBAs, and BAs to automate processing workflow that saved $2M in cost. • Constructed a Radar Transfer methodology using OpenCV to translate pixels to real-world coordinates that help find the violating vehicle. Optimized the algorithm further to differentiate Parked Vehicles. • Created an Automatic Frame Selection service using Computer vision to extract target vehicle image frames from video. • Built a Site Selection Model to find potential sites to set up new camera systems based on different key variables and deployed to Web using Flasgger. • Involved in research, design, and implementation of metrics monitoring that utilized anomaly detection techniques to create alerts and diagnose potential problems. Created and managed PowerBI reports for KPIs. • Developed and managed CI/CD pipelines using Azure to deploy automation services to different environments.