For two years, I worked as a research assistant at the University of Houston’s Structures and Artificial Intelligence Lab (SAIL), where I helped develop something with meaningful, real-world impact. Alongside Dr. Hoskere and a brilliant team, I co-developed TwInfra—an AI-powered system for automated 3D infrastructure assessment. What once took experts days—detecting corrosion, cracks, and structural issues—we brought down to under 30 minutes through deep learning, photogrammetry, and automated labeling. The system combined and expanded on two open-source projects, merging their strengths into a faster, more intelligent platform.

We built TwInfra using a full-stack pipeline of Docker, Python (Django, Flask, Celery), React, and Ubuntu, with Jupyter for experimentation and real-time data handling. I stayed active on the open repositories, contributing code, resolving issues, and helping shape the tools that powered our system. Though its applications reached into sensitive sectors, the purpose behind TwInfra was broader: to make critical inspections faster, smarter, and more accessible. The project even took me across the country, from presenting our work in Virginia to sharing it in Hawaii. These weren’t just conferences—they were reminders of why I build: to solve real problems with real stakes, and to push what’s possible through code.