During February 2025, Acherman developed per-class Intersection over Union (IoU) logging for the dsi-clinic/CMAP repository, focusing on enhancing model observability in computer vision workflows. Using Python and leveraging deep learning and data visualization skills, Acherman implemented a new metrics logging function and integrated it into both training and testing loops. This approach enabled granular, class-aware performance monitoring through TensorBoard, allowing for targeted insights and more efficient debugging. The work was locally verified using a debug workflow, with code cleanup pending. Acherman’s contribution provided a foundation for data-driven prioritization of improvements, reflecting a focused and technically sound engineering effort.