Lyceum · Zurich
You will work with the engineers who turn Lyceum’s GPU hardware into a platform customers can rely on. The team’s aim is a platform that operates without manual intervention: machines are provisioned, isolated and recovered automatically, and changes reach production without customers noticing.
You will own the layer that turns physical GPU servers into products customers can use: machines provisioned with the right stack, cleanly isolated between tenants, and kept running. Everything Lyceum sells, from GPU rental to VMs and inference, depends on this layer working.
Provisioning: Extend the orchestration layer that spins up and purposes GPU machines across GPU types, products and data centres, with the correct firmware, drivers, subnet manager version and OS image. Cut provisioning and cold-start times.
Isolation: Design and enforce tenant separation at network, storage and compute level, including clean wipe and re-provisioning between customers.
Uptime: Build health checks, automatic failure detection and remediation, so a failing node is caught and replaced before a customer notices. Take part in incident response.
Monitoring: Provide node- and fleet-level health and GPU telemetry, and make it available to the teams and customers who depend on it.
We consider candidates from diverse backgrounds, with a deep love for technical challenges and the desire to take on ownership beyond what's reasonably expected.
Requirements
Nice to have