During November 2024, Xiaoyue worked on the dsi-clinic/CMAP repository, focusing on improving the reliability of machine learning experiment management. Xiaoyue addressed a critical issue in the sweep configuration pipeline by implementing PyYAML-based parsing for the sweep_config.yml file, ensuring that training configurations are loaded accurately and consistently. This fix, developed in Python and leveraging configuration management skills, reduced the risk of misconfigurations during training sweeps and enhanced both reproducibility and traceability of experiments. The work demonstrated a targeted, in-depth approach to solving a specific problem, resulting in more robust experiment workflows and clearer audit trails for future development.