I am currently building research-oriented AI systems that connect computer vision models with practical robotic deployment.
My main direction is:
Designing lightweight perception frameworks that allow robots to understand complex environments using efficient visual AI models, geometric reasoning, and embedded deployment strategies.
This includes underwater pipeline inspection, bubble plume analysis, gas leak segmentation, UAV human detection, and monocular geolocation.
Lightweight Underwater Perception Framework NeptuNet is a research-oriented embedded AI framework for underwater gas pipeline inspection. Focus: Underwater Robotics · Embedded AI · Computer Vision · Gas Pipeline Inspection View Repository |
Lightweight UAV Perception and Geolocation System AlienSight is a UAV-based perception system for search-and-rescue scenarios. Focus: UAV Perception · Search and Rescue · Human Detection · Geolocation View Repository |


