Crusoe is on a mission to accelerate the abundance of energy and intelligence. As the only vertically integrated AI infrastructure company built from the ground up, we own and operate each layer of the stack — from electrons to tokens — to power the world's most ambitious AI workloads. When you join Crusoe, you join a team that is building the future, faster.

We're in the midst of the greatest industrial revolution of our time. The demand for AI compute is boundless, and power is a bottleneck. We're solving that — with an energy-first approach that makes AI infrastructure better for the world and faster for the people innovating with AI.

We're looking for problem-solving, opportunity-finding teammates with a sense of urgency, who believe in the scale of our ambition and thrive on a path not fully paved — people who want to grow their careers alongside a team of experts across energy, manufacturing, data center construction, and cloud services.

If you want to do the most meaningful work of your career, help our customers and partners advance their AI strategies, and be part of a high-performing team that believes in each other, come build with us at Crusoe.

There is no operating system for the AI datacenter — so we're building one.

Conductor is Crusoe's AI Infrastructure Operating Platform: a unified control plane that treats the entire global GPU fleet as a single programmable system. It is built on twin sources of truth — an observability data platform capturing runtime state from every GPU, NIC, switch, and host, and an inventory & topology graph modeling the intended state of every site, rack, cable, and component. Continuous reconciliation between the two, enforced by policy-gated workflows, drives the complete hardware lifecycle: from New Product Introduction through provisioning, validation, and burn-in to firmware upgrades, repair, and Day-2 operations — at the scale of 100,000+ GPUs.

This is a Principal Engineer role reporting into Cloud Availability, leading development of one of the highest-priority technical charters in the org. You will set the architecture, define the standards, and drive execution across the teams building the platform.

1. The platform architecture itself. Own the end-to-end design: twin sources of truth, the abstraction and orchestration layer (API gateway, workflow engine, policy engine, state reconciler), and the domain services built on top. You define how the pieces fit together and evolve.

1. The fleet as one logical computer. A topology graph that models every site, rack, host, GPU, switch, and cable as a traversable system of record — answering questions like "if I upgrade this spine switch, which customer workloads are affected?" in one query.

1. Reconciliation as the core loop. Never let the inventory system guess about runtime state, or telemetry guess about intended design. Detect drift between what should exist and what's actually running, and drive automated remediation — a GitOps-style loop for physical infrastructure.

1. Policy-gated lifecycle automation. No host reaches production without passing every gate: BOM validation, attestation, burn-in, network readiness, performance baselines. You design the policy engine that makes fleet growth a software-gated pipeline instead of a manual process.

1. The full hardware lifecycle as code. New Product Introduction, bare-metal provisioning, OS and firmware imaging, dependency-aware rolling firmware upgrades with canary stages and rollback, re-imaging, repair & RMA, and re-admission — orchestrated end to end, replayable via event sourcing.

1. Unified observability plane. One pane that correlates GPU, networking (InfiniBand/RoCE), storage, orchestration, and workload signals at extreme cardinality — so any engineer can diagnose and recover fast, and the platform can act on what it sees.

1. Site autonomy at global scale. Edge agents that keep every site operating through network partitions and reconcile with the global graph when connectivity returns — a distributed-systems problem across dozens of sites.

1. From detection to resolution, automatically. Anomaly detection, blast-radius computation from topology, workload migration, quarantine, RMA, burn-in re-validation, and fleet re-admission — closing the loop with no human in the path for known failure patterns.