Senior Software Engineer, AI Compute Scheduling
Core
Design and build scheduling systems for large GPU fleets across Slurm and Kubernetes to manage AI research workloads, including preemption, reservations, and fair sharing.
Role type
Senior IC software engineer (distributed systems/scheduling)
Builds
Production scheduling stack for GPU clusters, policy layers for resource allocation, and orchestration services.
Domain
High-performance computing (HPC) and AI infrastructure
Deliverable
production ML models | infrastructure
Required skills
Go, C/C++, Python, Kubernetes (operators, CRDs, scheduling), Linux systems, cluster schedulers (Slurm, Kueue, Volcano), distributed systems design, reliability engineering
Preferred skills
Large-scale AI training scheduling, NVLink/InfiniBand topology awareness, workflow orchestration (Flyte, Airflow, Ray), open-source contributions to scheduling projects
Technologies
Slurm, Kubernetes, Kueue, Go, C++, Python, Linux, Docker, cgroups, NVLink, InfiniBand, RoCE
Responsibilities
Design queueing and topology-aware placement for rack-scale systems; build policy layers for fair share and quotas; extend Kubernetes batch scheduling; develop reservation automation and node-health-aware placement; measure and optimize GPU goodput; own reliability and incident response; contribute to upstream open-source projects.
Seniority
Senior, hands-on IC