HPC Engineer

Verda is a Helsinki-headquartered full-stack AI cloud provider offering on-demand GPU compute, clusters, serverless containers, storage, APIs, and related AI infrastructure.

Helsinki, Finland
About Verda

Formerly DataCrunch, Verda operates European AI cloud infrastructure for the full AI lifecycle, from prototyping and distributed model training to scalable inference. It manages data-center infrastructure, hardware, cloud-platform services, and an in-house AI Lab.

View jobs by Verda

Skills

Candidate Availability

Required and preferred rules are kept separate and reflect the wording in the original posting.

About the Role

You will administer bare-metal and virtualized GPU and HPC clusters from provisioning through ongoing operations. You will design and tune InfiniBand fabrics, operate shared filesystems and workload scheduling, resolve hardware and software issues, maintain infrastructure records, and participate in on-call incident response.

Requirements

  • Linux skills
  • InfiniBand fabric design, subnet management, and performance tuning
  • IB clustering troubleshooting
  • Shared filesystem experience
  • Remote hardware diagnosis experience
  • NCCL, CUDA, DOCA, and NVIDIA stack knowledge
  • Slurm or workload scheduling knowledge
  • IPAM and DCIM record management
  • Production cluster operations experience
  • Python, Bash, or Ansible automation skills

Responsibilities

  • Administer bare-metal GPU and HPC clusters from provisioning through day-two operations
  • Administer virtualized clusters, including hypervisors and GPU virtualization
  • Design, deploy, and tune InfiniBand fabrics
  • Deploy and operate shared and parallel filesystems
  • Troubleshoot issues across fibers, transceivers, NICs, switches, drivers, and firmware
  • Coordinate with remote-hands and data-center teams on hardware fixes and expansions
  • Operate and tune Slurm or equivalent workload scheduling
  • Maintain issue tracking, IPAM, and DCIM records
  • Participate in on-call rotations and cluster incident response
  • Integrate new clusters with platform, network, and storage systems

Benefits

  • Equity compensation