Site Reliability Engineer (GPU Cloud)

About the Role

We are looking for a Site Reliability Engineer to support enterprise customers using GPU cloud infrastructure for AI inference workloads. This role blends software engineering and systems administration to design, monitor and scale large-scale GPU compute clusters.

What You'll Do

• Design the architecture of Kubernetes-based GPU container clusters, plan node pools, and implement network, storage and scheduling systems to ensure stable operation of both bare-metal GPU rental and containerized computing rental businesses.
• Implement GPU resource scheduling, multi-tenant isolation and fine-grained computing allocation based on NVIDIA GPU Operator, DCGM, MIG and other components, to deliver high-efficiency computing resources for AI training and inference.
• Develop automated scripts and operations tools in Python or Go to support platform workflows, including batch cluster deployment, tenant authorization, resource quota management, fault alerts and self-healing.
• Troubleshoot online cluster failures, conduct performance tuning and stability governance, and deliver technical solutions for cluster scaling, architecture optimization and risk prevention and control.
• Research mainstream cloud-native computing solutions in the industry, drive architecture iteration, and work with business teams to implement standardized computing power rental capabilities.

What You Bring

• Bachelor's degree or above in Computer Science, Software Engineering or a related field, with at least 3 years of experience in Kubernetes container cluster operation and maintenance.
• Hands-on experience in GPU computing cluster operation; familiar with GPU Operator, DCGM and containerized GPU scheduling.
• Proficient in Python or Go; able to independently develop automated scripts and operations tools.
• Holistic architectural thinking; able to deliver technical solutions for cluster deployment, scaling and migration.
• Expert in Linux system administration and the Prometheus / Grafana monitoring system.

Preferred Experience

• Prior knowledge of MIG, NUMA and NVLink
• DevOps development experience
• Experience operating clusters of hundreds of nodes
• Familiarity with distributed storage (Ceph, Lustre) and high-speed InfiniBand / RDMA networks

Work Location: In person