Note: The job is a remote job and is open to candidates in USA. CyberCoders is a company that builds high-performance, bare-metal GPU infrastructure for modern AI. They are seeking a Senior or Staff-level Platform Engineer to architect and operate GPU infrastructure, managing the full lifecycle of bare-metal GPU clusters and building automation for reliable AI workloads.
Responsibilities
- Design and operate container orchestration platforms optimized for NVIDIA DGX/HGX-class hardware
- Build bare-metal provisioning systems (PXE, Ironic, MAAS) to bring GPU clusters online at scale
- Manage GPU lifecycle: driver stacks, CUDA/kernel compatibility, MIG slicing, and performance tuning
- Partner with Network Engineering and DCOps to align physical infrastructure with software orchestration
- Build automation and internal tooling in Go or Python to streamline cluster operations
- Implement Terraform/Ansible-based IaC for fully auditable, repeatable infrastructure
- Design high-resolution observability stacks (Prometheus/Grafana, DCGM, VictoriaMetrics)
- Participate in a specialized on-call rotation supporting GPU workloads and core platform services
Skills
- 5+ years in Systems Engineering or HPC Infrastructure
- Strong Linux and bare-metal GPU experience
- NVIDIA DGX/HGX
- InfiniBand/RoCE
- Automation with Python or Go
- 7+ years in systems, platform, or distributed systems engineering (10+ for Staff)
- Expert-level Linux knowledge: kernel modules, sysctl tuning, hugepages, container runtimes
- Hands-on experience bootstrapping Kubernetes or SLURM on physical hardware
- Strong proficiency in Go (preferred) or Python for systems-level automation
- Deep familiarity with NVIDIA GPU ecosystems (drivers, CUDA, MIG)
- Working knowledge of InfiniBand or RoCEv2 networking and NCCL performance tuning
- Experience building observability pipelines for hardware-accelerated environments
- Ability to troubleshoot complex, multi-layered issues across hardware, networking, and orchestration
- Strong cross-team communication - you're the 'glue' between Network, DCOps, and Software
- Currently authorized to work in the United States without the need for sponsorship for a non-immigrant visa
- Experience with SLURM, Kubeflow, or distributed PyTorch
- Integrating vendor APIs (NetBox, Vault, GitLab CI, etc.) into unified workflows
- Infrastructure testing, chaos engineering, or cluster-level integration test suites
- Designing telemetry aggregation across hardware, networking, and environmental systems
Benefits
- Annual Bonus
- RSU's
- 5 weeks PTO
- 401k w/ match
- Comprehensive Benefit Plan
Company Overview