Worked on the NVIDIA/gpu-operator repository to enhance diagnostics and data collection for HyperConverged and KubeVirt environments within Kubernetes clusters. Developed a must-gather script extension in Shell that collects logs and YAML definitions for these resources, incorporating pre-checks to verify resource existence before gathering data. This approach reduced unnecessary data collection and minimized errors, improving the reliability of post-incident triage and root-cause analysis in complex, multi-tenant environments. Leveraged DevOps practices and Kubernetes expertise to deliver an end-to-end feature that strengthened observability and streamlined troubleshooting, reducing manual intervention during incident response and supporting reproducibility in cluster diagnostics workflows.