Staff Site Reliability Engineer – Bare Metal Linux – Data Center – Networking
Location: 100% remote – US
Compensation: up to $350k OTE
Our client, a publicly listed AI cloud provider, is building a platform for high-throughput, compute-heavy workloads. They run 1,000+ bare-metal servers across their own datacenters, where failure modes are real, capacity is finite, and reliability needs to be engineered, not “handled”.
We’re looking for a Senior/Staff SRE to own production reliability end to end. You’ll define SLIs/SLOs, run error-budget conversations, and ship changes that reduce incidents and improve latency (p95/p99). You’ll build automation to kill toil, make deployments safer (canary/rollback), and turn observability into signal rather than noise.
This is a bare-metal environment: Linux, datacenters, physical fleets and real hardware constraints, not managed services. You’ll work close to the metal across Linux performance (CPU/memory/I/O contention), network debugging (DNS/TCP/TLS, packet loss, congestion) and Kubernetes running on bare metal. On-call is part of the job, but success is measured by how much you reduce it.
Must-haves
- Extensive, recent production experience running bare-metal / on-prem / datacenter infrastructure (not public cloud only), typically 15+ years in Linux-heavy roles
- Experience operating large fleets of physical servers, directly or as part of a team
- Deep hands-on Linux debugging and performance work at kernel level (CPU, memory, I/O, low-level behavior)
- Strong networking: DNS/TCP/TLS, latency, packet loss, congestion, troubleshooting under load
- Strong coding in Go and/or Python, beyond automation scripts. Real engineering capability is a must.
- Experience in low-latency environments, e.g. kernel tuning, CPU pinning/NUMA, kernel bypass, RDMA/InfiniBand, HFT or GPU cluster networking
Nice to have
- Kubernetes internals on bare metal: scheduler behavior, autoscaling edge cases, kubelet pressure/evictions, etcd/control plane
- InfiniBand
- C development
- Terraform, Puppet or Ansible; Docker, Helm and modern CI/CD
Why join
- Work on large-scale infrastructure powering the AI economy
- 100% remote, with competitive compensation
If you’re looking for complexity and a new place to nerd out on infrastructure optimization, we’d love to hear from you!