Research Assistant
Current
University Of Houston
• Establishing interoperability between open-source computer vision software (ODM & CVAT), enabling asynchronous 3D model generation via photogrammetry and automatic labeling of infrastructure defects like cracks and corrosion on naval ships with deep learning models.• Architected a serverless function via Nuclio to host a deep learning model designed for precise corrosion detection. • Leveraging capabilities in several Deep Learning models, Docker, Python frameworks (Django, Flask, Celery), Jupyter, and API usage with Requests and HTTPS along with Linux (Ubuntu) proficiency for development.• Contributed to both the GitHub repositories and community forums for OpenDroneMap (ODM) and Computer Vision Annotation Tool (CVAT), committing code, participating discussions, proposing 2 improvements, and providing support to over 7 other users and contributors.