Role: 

• Designed and deployed Edge AI computer vision solutions for indoor and outdoor smart parking management systems.

• Developed multi-object detection models achieving 99% accuracy across diverse parking environments.

• Built Fire and Smoke detection models with 99% accuracy for early hazard detection, enhancing EV fire safety in parking and office environments.

• Implemented AI algorithms to accurately estimate vehicle ground center points using Detect View–Aerial View 2D board mapping.

• Conducted applied research to develop customized detection pipelines for varying parking layouts and operational conditions.

• Performed advanced research on LPR, OCR, 3D Bounding Boxes (3D BB), Oriented Bounding Boxes (OBB), 3D Segmentation, DepthPro, Vision-Language Models (VLM), and Multimodal LLMs (MM-LLM) for real-world parking scenarios.

• Designed and maintained CI/CD pipelines using GitHub Actions, integrating automated testing in C++.

• Led data collection, annotation, and model experimentation, managing datasets and experiments with DVC over SSH/WebDAV.

• Developed an automated model analysis and testing framework with hourly, daily, and weekly triggers, consolidating field-test results and significantly improving model validation efficiency and team productivity. 

Project Title:   Development of Environment Adaptive Video Surveillance System using Intelligent Management of Swarm Robots

Project Title:   Development of Environment Adaptive Video Surveillance System using Intelligent Management of Swarm Robots

Department of Computer Science and Engineering,

Narasus Sarathy Institute of Technology, Tamil Nadu, India.

DURATION: (Jun-2013 ~ May-2016)