maneki.
POST
MANEKI/JOBS/FORESIGHT INSTITUTE
Foresight Institute
Foresight Institute
6 open roles on Maneki

Compute Support & Systems Administrator (Part-Time Contractor - Berlin)

Entryon site
LOCATION
—
POSTED
1w ago
APPLY DIRECTLY
Application goes to Foresight Institute's own career page
APPLY ON FORESIGHT →
Maneki aggregates jobs from top Web3 companies. We redirect you to the original posting — no middleman, no added steps.

Required skills

UnixInternetworkingServer Hardening

Job description

Compute Support & Systems Administrator

Part-Time Contractor — Berlin

Foresight Institute is hiring a hands-on Systems Administrator to support the technical infrastructure of our Berlin AI Node — a physical hub where researchers collaborate on AI-driven scientific progress.

Are you a practical, implementation-focused person with an entrepreneurial spirit? At the Node you will help keep our local compute infrastructure running smoothly, support containerized workloads, and troubleshoot issues when they arise. Some work is remote, but you must be available for on-site hardware setup and occasional urgent issues in the Mission, Berlin.

Role details

  • Location: Berlin (CIC, Lohmühlenstraße 65, 12435 Berlin, Germany)
  • Type: Part-time contractor, ad-hoc (2–10 hours/week)
  • Compensation: per hour (based on experience)
  • Start: ASAP (rolling review)

What you’ll do

  • Maintain and troubleshoot UNIX-based GPU compute systems
  • Maintain local servers and physical hardware
  • Support Docker/containerized research workloads
  • Manage NVIDIA drivers, CUDA environments, and multi-user GPU access
  • Help onboard research projects and improve pre-existing documentation

Required

  • Formal and informal hands-on experience with physical servers or local hardware
  • Experience with NVIDIA GPUs (drivers, CUDA, troubleshooting)
  • Able to diagnose infrastructure and networking issues
  • Based in Berlin and able to come on-site when needed
  • Experience with Docker and Unix-based systems
  • Ability to diagnose networking and hardware issues
  • Reliable and responsive (typically within 16 hours)

Nice to Have

  • Experience with multi-user GPU scheduling (e.g., Slurm, Kubernetes, or similar)
  • Familiarity with ML research workflows (PyTorch, JAX, etc.)
  • Experience in research or academic compute environments
  • Basic security hardening knowledge
  • Experience with heterogeneous or mixed-architecture clusters

If you enjoy enabling cutting-edge research by making infrastructure work smoothly, we’d love to hear from you. Apply here.