Software Infrastructure Engineer

C5R CORP; San Francisco (CA); On-site; USD 150,000 - 210,000; Full time

Job summary

C5R Corp builds research facilities that run frontier AI experiments, connecting thousands of devices and storing petabytes of data. You will design, build, evolve, and maintain robust infrastructure to manage the lab environment and data workflows.

Join a team responsible for networking, scheduling workloads, and long-term data storage critical to validating AI experiments and scientific results.

Qualifications

• Experience designing robust infrastructure for lab-scale systems.
• Ability to provision and manage networked devices and services in a heterogeneous environment.
• Track record of building reliable data storage for telemetry, logs, and results.

Responsibilities

• Design, build, evolve, and maintain routing and switching infrastructure for a lab with thousands of devices.
• Develop and maintain a custom job scheduler to run programs across diverse machines.
• Ensure reliable time-series data storage for cameras, logs, telemetry, and research outputs.

Skills

Routing infra

Job scheduler

Data storage

Tools

Time-series database

Job description

C5R builds research facilities where frontier AI models design experiments, operate instruments, and learn from real scientific work.

Our facilities run on hundreds of switches, thousands of devices (H100s, Orins, custom-built USB-Ethernet dongles, MacBook Neos), and petabytes of data. We need engineers to design and build simple, robust infrastructure to manage all of it. You will design, build, evolve, and maintain systems like these:

• Routing and switching infrastructure: every component in the lab is on the network. Keeping thousands of heterogeneous devices provisioned, addressable, and reliable is a core engineering problem, not an afterthought.
• Job Scheduler: we run a diverse set of programs across a diverse set of machines, scheduled by a process manager we built ourselves.
• Data storage: every aspect of the lab is recorded (cameras, software logs, hardware telemetry, equipment output, research results) and stored in a time-series database. This data verifies the accuracy of our results and trains frontier models for science.

See full description here: https://c5r.net/jobs/sie

See my thoughts on how we use AI to develop infrastructure here: https://logofry.com/golden-age