Developed an enhancement for the timholy/boltz repository focused on improving experiment tracking for machine learning workflows. Introduced a Run Name parameter to the Weights & Biases logger, enabling more precise identification and organization of training runs within the platform’s dashboards. This addition supports better observability, reproducibility, and data governance by allowing users to distinguish between experiments efficiently. The work was implemented in Python and leveraged integration with Weights & Biases, following version-controlled development practices. No major bugs were addressed during this period, with efforts concentrated on feature delivery to streamline experiment management and reduce time spent on debugging and reporting.