Supercomputing Engineer, Fleet Software

Posted: 2025-06-18

About Etched

Etched is building hardware for frontier intelligence. We co-design chips, racks, software, and manufacturing to deliver best-in-class throughput and latency across both prefill and decode workloads. Our first products are heavily focused on inference. Backed by hundreds of millions from top-tier investors and staffed by leading engineers, Etched is redefining the infrastructure layer for the fastest growing industry in history.

Fleet Software enables deployments at scale through configuration management, safe software and firmware rollouts, and software recovery, while partnering with other engineering teams to drive product improvements as we ramp production at unprecedented speed. The team works closely with Node Systems, which builds the single-rack software, clients, and tuning recipes that Fleet deploys and manages across many racks.

The team also builds the monitoring and analysis systems that connect signals across the rack. Engineers and their agents use this data to investigate failures, spot patterns, and continually improve the performance and reliability of the inference fleet. We're looking for a hands-on engineer who has built infrastructure and owned it through real deployments in the field.

- Build the fleet management system that provisions, configures, updates, and recovers Etched-owned hardware across labs and customer deployments

- Implement safe software and firmware rollouts, including validation, staged deployment, and handling of partial failures

- Build telemetry and log pipelines that collect rack metrics and logs into a central store, including decisions on storage, retention, and query performance

- Develop monitoring, alerting, and analysis tools that help engineers, their agents, and customers understand fleet behavior and investigate failures

- Build secure, least-privilege access to BMCs and servers across lab, office, and data center networks

- Give lab operators simple controls for managed hardware, such as draining and powering down for repair, and rolling out validated versions

- Partner with Node Systems to define the interfaces, clients, and validated configurations needed to manage hardware across the fleet