Casey Kinsey

Architected and built Decker, a SaaS platform for UAV-based bridge deck inspection for state DOTs. Owned full-stack development from raster ingestion through model inference to the end-user web app. Designed and shipped computer-vision architectures for automated defect detection in infrared imagery, including training pipelines, model serving, and statistical confidence reporting for delamination assessment. Built geospatial processing pipelines for large raster handling, CRS normalization, tiling for web delivery, and integration with field-collected UAV survey data. Drove practical adoption of LLMs inside engineering through internal copilots, customer-facing AI features, and evaluation harnesses; productized from prototype to a platform powering 12 state DOT pilots.