Developed and integrated Vertical Pod Autoscaler (VPA) support for the NVIDIA/KAI-Scheduler project, focusing on enhancing resource efficiency and automation within Kubernetes environments. The work involved implementing VPA specifications directly into configuration files and updating core components such as admission and binder to utilize these dynamic resource settings. By enabling configuration-driven VPA management, the changes allow KAI-Scheduler workloads to benefit from automated resource optimization and improved cluster utilization. The solution was delivered using Go and leveraged expertise in cloud infrastructure and Kubernetes, aligning the scheduler with autoscaling best practices and laying the foundation for more adaptive and efficient workload management.