Mark developed a robust vision-based pose estimation system for the gatorbotics1700/Robot_2025 repository, focusing on improving autonomous alignment and field-relative positioning. He implemented AprilTag-based pose calculation within the LimelightSubsystem, replacing legacy logic with a transforms-based approach to accurately convert camera-relative poses to field space. Using Java and leveraging skills in computer vision and coordinate transformations, Mark refactored control flow to centralize pose computation and introduced comprehensive unit tests with JUnit to validate the new logic. This work enhanced code quality and maintainability, providing a solid foundation for reliable robot localization and ongoing regression safety in robotics applications.