Senior Cluster Orchestration Engineer
Core
Build a shared GPU cluster layer that unifies siloed capacity for training, inference, and batch workloads on Kubernetes and SLURM.
Role type
Senior IC platform/infrastructure engineer (GPU cluster orchestration)
Builds
Shared GPU capacity platform, schedulers, controllers, and platform services
Domain
AI Infrastructure / Cloud Computing
Deliverable
production ML models | infrastructure
Required skills
Go or Python, Kubernetes (scheduler, control plane, operators), GPU workload scheduling, batch/AI schedulers (Volcano, KAI Scheduler, Kueue, Slurm Bridge), gang scheduling, topology-aware placement, preemption, observability, GitOps
Preferred skills
Experience with perishable compute, elastic training, checkpoint-aware workloads, SLURM vs cloud-native scheduling models
Technologies
Kubernetes, SLURM, Volcano, KAI Scheduler, Kueue, Slurm Bridge, GitOps
Responsibilities
Design scheduling layer for shared GPU capacity; Build support for perishable compute; Implement gang scheduling and preemption; Develop Kubernetes operators for quotas and lifecycle; Solve workload interactions (elastic training, inference autoscaling, batch backfill); Build observability for scheduling and utilization; Deliver platform via GitOps
Seniority
Senior, hands-on IC
