Senior Computer Vision Researcher
Current• Built a multi-vehicle tracking system in pytorch from multiples camera angles on school buses with license plate detection and reading using yolov8, strongsort, classical AI search optimzation algorithms and post image processing. Obtained 95% accuracy.• Using off the shelf latest AI algorithms to detect make and color of cars, monocular depth estimation, projecting one camera view unto another using homography projection and lane detection.• Performed camera calibration for scene understanding and 3D reconstruction• Deployed and ran AI solutions on the Nvidia Orin edge device using DeepStream• Setup a CI/CD ML testing platform with Docker, CML and Github Actions, on on-premise GPUs to manage ML experiments in order to validate multiple models before deployment.