Worked on the kubernetes/kubernetes repository to address a critical bug in the scheduler’s preemption logic, focusing on improving correctness and reliability for high-priority workloads. Using Go and Kubernetes, implemented deterministic sorting to ensure that higher-priority pods are not incorrectly evicted during preemption events. Enhanced the scheduler’s victim selection process by updating the sorting mechanism and expanded the test suite to validate behavior across various priority scenarios. This work improved resource utilization and quality of service for critical pods, reduced the risk of disruptive preemption, and strengthened overall cluster reliability, demonstrating strong skills in Go, Kubernetes internals, and automated testing.