Computer Vision & Machine Learning Engineer

The University of Wisconsin–Milwaukee’s College of Engineering & Applied Science is an ABET accredited hub for hands-on engineering and computer science education, blending interdisciplinary research with strong industry partnerships.

• Drove cross-disciplinary collaboration with civil-engineering faculty and PhD scholars, translating CV derived GPS trajectories into roadway-risk metrics and design recommendations; co-authored two manuscripts that formalize the project’s safety impact and data methodology.
• Led development of a real-time monocular 3D perception pipeline (YOLOv9, DeepSORT, custom GPS DNN) for 8K panoramic footage, streaming 30 FPS with 94% mAP and 98% pixel-to-GPS accuracy.

Software Engineer (R&D)

mFilterIt provides data-driven solutions to ensure transparency, safety, and authenticity in the digital ecosystem by detecting invalid traffic, ad fraud, and brand safety risks across digital platforms.

• Worked on the Brand Safety and Creative Compliance Moderation projects to ensure ad placements met brand standards and government regulations, using automated systems to filter harmful content and enforce policy compliance.
• Developed an industrial-scale fraud detection pipeline to identify brand misuse across platforms like Quora, Facebook, and Twitter, leveraging YOLOv5, OpenCV, Docker, and REST APIs on Ubuntu with Python.

Research and Development Engineer- Machine Learning & Computer Vision

VaaaN is a leading provider of intelligent traffic solutions, focused on enhancing transportation infrastructure through advanced, customizable technologies. Since its inception in 2011, the company has grown rapidly in India, offering end-to-end turnkey services from design to deployment and maintenance for smarter road and traffic management.

• Developed and deployed a camera-based Automatic Vehicle Classification (AVC) system using computer vision and 180-degree fisheye cameras on NVIDIA Jetson Xavier, achieving real-time toll automation with 97% true positive accuracy through YOLOv3, TensorRT, and TensorFlow 2.0.
• Secured a national patent (Patent No. 538015) in collaboration with VaaaN Infra Pvt. Ltd. for the AVC system, now actively used in India’s transportation infrastructure.

Computer Vision and Machine Learning Intern

• Worked on object detection and classification using YOLO and Convolutional Neural Networks (CNN) to identify and localize multiple object classes in visual data. Optimized model performance for real-time inference in various detection scenarios.
• Implemented HOG-based vehicle classification using LiDAR datasets to extract geometric features and classify vehicle types. Combined traditional feature engineering with point cloud processing for enhanced accuracy in 3D object recognition.