Systems Engineer / Site Reliability Engineer (SRE)

Bellatrix Systems LLC; San Jose (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

Bellatrix Systems LLC in San Jose, CA invites an experienced Systems Engineer / SRE to design, automate, and maintain large-scale AI inference infrastructure. You will focus on GPU cluster reliability, bare-metal provisioning, Kubernetes orchestration, and Linux performance optimization to support distributed accelerator environments.

The role emphasizes high availability, scalability, and performance. You will implement monitoring, telemetry, automated recovery, and root-cause analysis to

Qualifications

• Experience in hyperscale or GPU cloud infrastructure.
• Experience with NVIDIA GPU drivers, CUDA, Fabric Managers, NCCL.
• Kubernetes cluster management at scale.
• Programming in Python or Go.
• GPU networking: RoCEv2/InfiniBand.
• SLIs/SLOs, telemetry, automated self-healing, root-cause analysis.

Responsibilities

• Design and maintain scalable GPU infrastructure for AI inference workloads.
• Automate bare-metal provisioning and server lifecycle management.
• Manage Kubernetes clusters and optimize Linux/GPU performance.
• Troubleshoot GPU, network, and distributed system bottlenecks.
• Implement monitoring, telemetry, and automated recovery solutions.
• Conduct root-cause analysis and improve infrastructure reliability.

Skills

Hyperscale infra

Distributed systems

NVIDIA drivers

CUDA

Fabric Managers

NCCL

Kubernetes at scale

Python

Go

RoCEv2

InfiniBand

SLIs/SLOs

Telemetry

Self-healing

Root-cause

Job description

Duration: Long term

Job Summary

Seeking a Systems Engineer / SRE to design, automate, and maintain large-scale AI inference infrastructure. The role focuses on GPU cluster reliability, bare-metal provisioning, Kubernetes orchestration, Linux performance optimization, and infrastructure automation. Responsible for ensuring high availability, scalability, and performance across distributed accelerator environments.

Must-Have Skills

• Experience in hyperscale environments, distributed systems, or GPU cloud infrastructure.
• NVIDIA GPU drivers, CUDA, Fabric Managers, NCCL.
• Kubernetes cluster management at scale.
• Programming: Python or Go.
• GPU networking: RoCEv2, InfiniBand.
• SLIs/SLOs, telemetry, automated self-healing, root-cause analysis.

Responsibilities

• Design and maintain scalable GPU infrastructure for AI inference workloads.
• Automate bare-metal provisioning and server lifecycle management.
• Manage Kubernetes clusters and optimize Linux/GPU performance.
• Troubleshoot GPU, network, and distributed system bottlenecks.
• Implement monitoring, telemetry, and automated recovery solutions.
• Conduct root-cause analysis and improve infrastructure reliability.