Worked on the OpenHUTB/nn repository to integrate a CARLA-based simulation module, focusing on improving the development workflow for computer vision and machine learning tasks. Delivered a new testing workflow using Python, clarified environment setup, and standardized file and folder naming conventions to enhance maintainability and reduce setup friction. Updated documentation, including localizing the README to Chinese, to support onboarding for a broader contributor base and minimize support overhead. The work emphasized code hygiene and streamlined simulation-driven feature validation, laying a foundation for faster iteration cycles and more reliable data processing within the repository’s evolving machine learning infrastructure.