Andromeda
Member of the Technical Staff - Platform
North America Remote / San Francisco, CA · Remote · Staff · Full-time · Posted Jul 30, 2026
- Experience
- Impressive technical work with impact; the posting says this may represent three to twenty years of experience.
About the job
Build and operate Andromeda’s bare-metal Kubernetes platform, making GPU clusters reliable and able to serve workloads across providers, sites, and hardware generations.
What you'll do
- Detect and repair failing GPUs, links, and hosts; manage node setup, wiping, verification, and customer handoffs.
- Build workload scheduling, virtualization, storage, and network and storage isolation systems.
- Run the platform in production, respond to on-call incidents, and scale its reliability.
What they ask for
- Have written Kubernetes controllers or operators.
- Have debugged hardware and kernel failures in production on Linux and bare metal.
- Have shipped production services in Go, Rust, or C, and be able to ship production Go within the first month.
- Have production on-call experience and demonstrated ownership of production systems.
Nice to have
- Experience with HPC schedulers such as Slurm, virtualization, or distributed storage.
- Experience with high-performance networking or out-of-band management.
Tools
Kubernetes, Go, Rust, C, Slurm, KVM, QEMU, Ceph
AI focus
Build and operate infrastructure for GPU clusters used by AI labs and model-training workloads.
Model-written summary · the original posting is authoritative