Worked on enhancing the reliability of the grafana/scheduler-plugins repository by developing a comprehensive testing suite for the PEAKS Scheduler Plugin. Focused on quality assurance, this work involved creating both unit and integration tests in Go to validate the plugin’s initialization, scoring, and normalization logic within the Kubernetes scheduler. The tests specifically addressed node score normalization based on power models and watcher metrics, ensuring correct behavior in production environments. By improving test coverage and automating validation of core scheduling logic, the changes reduced regression risk, enabled safer pull request reviews, and contributed to faster, more reliable release cycles for the project.