Worked on the NVIDIA/gpu-operator repository to enhance deployment reliability by addressing a namespace scoping issue in the upgrade_crd.yaml configuration. Focused on DevOps practices, the developer identified and fixed a missing ServiceAccount namespace, ensuring that deployments using ArgoCD and Helmfile are correctly scoped and reducing the risk of upgrade failures across Kubernetes clusters. The solution involved precise YAML configuration changes and demonstrated a strong understanding of Kubernetes RBAC and Git-based workflows. This targeted bug fix improved consistency and reliability in multi-environment deployments, reflecting careful attention to detail and a methodical approach to infrastructure automation and cluster management.