Worked on the NVIDIA/doca-platform repository to deliver a DMS Pod Scheduling Enhancement that improved deployment reliability for DMS workloads. Focused on backend development using Go, the solution relaxed taint tolerations by leveraging the Exists operator in Kubernetes, allowing DMS pods to be scheduled on nodes with any taint. This approach reduced the need for manual taint management and increased scheduling flexibility, making the platform more resilient to changes in node taints. The enhancement contributed to better resource utilization and cluster reliability, demonstrating effective use of cloud computing concepts and Kubernetes primitives to address operational challenges in workload scheduling.