Technical Lead - GPU Infrastructure (100% Remote - Worldwide)

Cosmic AC is Tether Data's GPU compute and managed inference platform. The platform is expanding to own the full stack on bare-metal GPU infrastructure: a managed Slurm scheduling layer for internal research/model-training teams, then a Kubernetes control plane for inference tenancy. The Technical Lead owns architecture and delivery, and line-manages roughly 12 engineers across Europe and India.

Responsibilities

• Own end-to-end platform architecture and lead/line-manage the distributed team
• Design, build, and operate a managed Slurm service, including partitions, node onboarding, and health/autohealing
• Own Kubernetes cluster bootstrap on partner bare metal using the NVIDIA GPU Operator and KubeVirt/VFIO GPU isolation
• Build managed inference at scale, including multi-GPU/multi-node serving and autoscaling
• Own observability, SLOs, and incident response; interface with infrastructure vendors on capacity planning

What we are looking for

• 8+ years hands-on engineering including 3+ years leading infrastructure platform teams
• Hands-on Slurm at scale (slurmctld/slurmdbd, partitions, QoS, node health scripting)
• GPU fleet operations on bare metal: NVIDIA driver/CUDA lifecycle, Fabric Manager/NVSwitch, DCGM, MIG
• High-performance interconnects: InfiniBand, RDMA, SR-IOV, NCCL debugging

Additional Information

100% remote, worldwide, managing a distributed engineering team spanning Europe and India.