Staff HPC Network Engineer

Atoms•San Francisco, CA

2d•$224,000 - $284,000•Onsite

About The Position

Atoms is seeking an HPC Network Engineer to join their founding team. This role will be responsible for designing and running the high-performance network that connects the company's GPU compute resources. The engineer will focus on building a fast, reliable fabric and developing the architecture to scale it as the cluster grows. This position is based in San Francisco and requires onsite work five days a week.

Requirements

Experience designing, deploying, and scaling networks for HPC or GPU compute environments.
Hands-on experience managing switches across platforms such as Arista (EOS), Cisco (NX-OS), Nvidia (Cumulus), and/or SONiC.
Solid capacity-planning and project-management skills.
Packet decoding/analysis (tcpdump, Wireshark).
Knowledge of fiber optics and high-speed cabling up to 800 Gbps.
Familiarity with public cloud networks (AWS, GCP, Azure).
Familiarity with config management (Ansible, Salt).
Familiarity with Python, Prometheus, Grafana, ELK, and GitHub.
Awareness of data center power and cooling.
Meticulous attention to detail, proactive problem-solving, and excellent communication.

Responsibilities

Design, optimize, and scale the high-performance network fabric that interconnects GPU and CPU compute.
Build low-latency networking to ensure training and simulation workloads are efficient and fast.
Own capacity planning and lead network projects from design through deployment.
Partner with vendors to select and engineer switches for future compute growth.
Develop scripts and processes to automate network operations and improve efficiency.
Collaborate with the infrastructure team to solve cross-discipline problems with immediate operational impact.