Led an international student team in implementing the backend system for AIOT safety identification technology improvement plan for structural steel construction sites.
Completed the system one year ahead of schedule. Supervised 2 construction sites concurrently in real-time using AI, highlighting effective coordination and successful project execution.
Implemented the Django framework, utilized Celery and Redis for asynchronous task processing, integrated MySQL for data storage, employed RESTful API for microservices such as AIOT integration and Line Bot notification, and utilized Docker/Git for containerization and version control.
Annotated data and trained YoloV4 and U-Net model to identify safety equipment and hazardous behaviors, achieving 94% precision and IOU of 77.68%.