Andromeda
Senior Site Reliability Engineer
Global Remote / San Francisco, CA · Remote · Senior · Full-time · Posted Nov 4, 2026
About the job
Designs, operates, and troubleshoots large-scale GPU infrastructure for distributed AI training and inference, including customer support, reliability, performance, and cluster automation.
What you'll do
- Design and operate multi-provider, multi-region GPU clusters, including scheduling, networking, storage, and capacity planning.
- Partner with customers to onboard, troubleshoot, and optimize large-scale training workloads; lead incident response and postmortems.
- Build GPU-specific observability and production automation for provisioning, health checks, scheduling, self-healing, and lifecycle management.
What they ask for
- Hands-on production experience operating large-scale GPU clusters (NVIDIA A100/H100/B200 or equivalent).
- Production experience with InfiniBand, RoCE, or NVLink fabrics for distributed training.
- Expert-level Linux systems knowledge and strong experience running Kubernetes in production with GPU workloads; Slurm or another HPC scheduler is also valued.
- Strong engineering skills in Python, Go, or Bash, and Infrastructure-as-Code proficiency with Terraform, Helm, Ansible, or equivalent.
Nice to have
- Experience with high-performance parallel file systems such as VAST, Weka, Lustre, or GPFS, or optimizing distributed training performance.
- Experience with physical cluster design, hardware validation, or leading and mentoring infrastructure engineers.
Tools
NVIDIA A100/H100/B200, InfiniBand, RoCE, NVLink, NCCL, CUDA, PyTorch, Kubernetes
AI focus
Operates and optimizes GPU infrastructure used for large-scale AI model training and inference.
Model-written summary · the original posting is authoritative