Worked on the wandb/helm-charts repository to deliver resource allocation optimizations for the MCP-Server component, focusing on improving cluster efficiency and reliability. The approach involved right-sizing Kubernetes resource requests by reducing CPU from 500m to 50m and memory from 1Gi to 512Mi, aligning resource reservations with actual idle utilization. Updated Helm chart configurations and corresponding test snapshots to ensure deployment reliability after these changes. Leveraged skills in Helm, Kubernetes, and Infrastructure as Code to implement and validate the optimizations. No major bugs were reported during this period, reflecting a targeted and stable engineering effort focused on infrastructure resource management.