Worked on the roboflow/supervision repository to optimize heatmap generation in the HeatMapAnnotator module using Python and computer vision techniques. Refactored the HSV color mapping process by applying the temperature map directly to the hue channel while maintaining maximum saturation and value, which eliminated redundant HSV conversions and improved performance. This approach resulted in a 20x speedup, reducing CPU usage and enabling more scalable annotation workflows. Focused on performance optimization and maintainability, the work included profiling and commit-based traceability to document improvements. No major bugs were addressed during this period, with efforts concentrated on feature development and code efficiency.