Rajat Choudhary contributed to the NVIDIA/gpu-operator repository by developing a sandbox device plugin that enhances GPU feature discovery within Kubernetes clusters. Leveraging containerization and YAML configuration, Rajat enabled the GPU Feature Discovery (GFD) pods to launch with the necessary privileges and environment variables, improving visibility into GPU capabilities. Additionally, Rajat addressed operational resilience by fixing device plugin permissions, granting update rights to NodeFeatures and ensuring stable functionality during node restarts. This work improved the reliability of GPU workload scheduling and reduced downtime during maintenance, demonstrating a solid understanding of Kubernetes device plugin architecture and DevOps best practices.