Worked on the pytorch/ignite repository to modernize type hints for the ROC AUC metric, focusing on aligning the codebase with Python 3.10+ standards. Leveraging expertise in Python and type hinting, updated the ignite/metrics/roc_auc.py module to use contemporary syntax, which enhances code clarity and reduces the risk of typing-related regressions. This targeted feature improvement did not involve bug fixes but laid the groundwork for broader typing modernization across the metrics module. The work supports maintainers and users transitioning to newer Python versions, ensuring ongoing compatibility and maintainability within the software development process for machine learning evaluation metrics.