Job Summary
We are seeking an experienced Systems Engineer / Site Reliability Engineer (SRE) to design, automate, and operate the core infrastructure platform supporting high-density AI inference environments.
In this role, you will work at the intersection of bare-metal accelerator hardware, Linux systems engineering, Kubernetes orchestration, and large-scale distributed infrastructure. You will be responsible for the reliability, scalability, performance optimization, and lifecycle automation of compute infrastructure supporting large-scale GPU workloads.
The ideal candidate will have extensive experience in Linux systems engineering, hyperscale infrastructure, GPU computing environments, and production-grade automation. You should have hands-on expertise with NVIDIA drivers, CUDA, Kubernetes, infrastructure as code, and automated provisioning technologies.
Key Responsibilities
- Design, build, and maintain automated infrastructure provisioning pipelines using PXE, iPXE, and Zero Touch Provisioning (ZTP).
- Manage and optimize Kubernetes clusters supporting large-scale GPU and AI inference workloads.
- Perform advanced Linux systems administration, kernel tuning, and performance optimization.
- Troubleshoot Linux kernel, systemd, NUMA, PCIe, and storage I/O performance issues.
- Manage GPU infrastructure, including NVIDIA drivers, CUDA, NCCL, and fabric managers.
- Develop infrastructure automation using Python or Go.
- Implement configuration management and infrastructure as code using Terraform, Ansible, and SaltStack.
- Establish and maintain SLIs, SLOs, monitoring, and observability frameworks.
- Build automated self-healing mechanisms to improve infrastructure reliability and availability.
- Troubleshoot distributed systems, hardware, networking, and GPU performance bottlenecks.
- Drive root cause analysis (RCA) and implement permanent solutions to recurring infrastructure issues.
- Improve operational efficiency through automation and reduction of manual intervention.
- Support the lifecycle management of large-scale compute and accelerator infrastructure.
Required Qualifications
- Strong experience as a Systems Engineer, Site Reliability Engineer, or Infrastructure Engineer.
- Extensive hands-on experience with Linux internals, including kernel tuning, systemd, and performance troubleshooting.
- Strong experience managing Kubernetes clusters in production environments.
- Experience with bare-metal provisioning technologies such as PXE, iPXE, and ZTP.
- Hands-on experience with NVIDIA GPU infrastructure, drivers, CUDA, and NCCL.
- Strong understanding of NUMA architectures, PCIe subsystems, and storage I/O optimization.
- Proficiency in Python or Go for infrastructure automation and tooling.
- Experience with Terraform, Ansible, or SaltStack.
- Knowledge of monitoring, telemetry, SLIs, SLOs, and infrastructure observability.
- Experience supporting large-scale distributed systems and hyperscale infrastructure.
- Strong troubleshooting, debugging, and root cause analysis skills.
Preferred Qualifications
- Experience working with GPU cloud providers, AI infrastructure companies, or hyperscale cloud environments.
- Previous experience at companies such as CoreWeave, Nebius, Crusoe Cloud, AWS, or Google Cloud.
- Experience with high-performance computing (HPC) and distributed GPU clusters.
- Familiarity with Prometheus, Grafana, VictoriaMetrics, and OpenTelemetry.
- Understanding of high-performance networking technologies, including RoCEv2 and InfiniBand.
- Experience with automated infrastructure recovery and self-healing systems.
- Experience operating infrastructure across thousands of bare-metal servers or accelerator nodes.
Ideal Candidate Profile
We are looking for a hands-on infrastructure engineer with deep Linux systems expertise and strong SRE fundamentals. The ideal candidate should understand how bare-metal hardware, GPU accelerators, operating systems, and Kubernetes work together to support large-scale AI inference workloads.
Candidates with experience operating GPU clusters, hyperscale compute infrastructure, and highly automated data center environments will be strongly preferred.