At NVIDIA, I deliver production infrastructure across FleetCommand and Dynamo, spanning resilient GPU edge operations and latency-sensitive distributed inference. I redesigned FleetCommand deployments to make remote GPU rollouts durable, retry-safe, and recoverable across intermittent connectivity.
At Red Hat, I productized Kubernetes-native virtualization for OpenShift and shaped update architecture for disconnected clusters. Earlier, I defined kolla-kubernetes architecture and developed OpenShift automation workflows, while working on OpenStack containerization and SELinux fixes.

