Director of Engineering, Flex Compute
Crusoe · San Francisco, CA - US
About this role
**Director of Engineering, Flex Compute** **About the Role:** Lead the design and development of Crusoe's Flexible Compute system—a utility-facing, safety-critical control system that manages AI infrastructure power delivery. You'll own the curtailment orchestration layer, ship the first production system, and build the team from scratch. **Key Responsibilities:** • Own curtailment orchestration and decision layer: grid-signal ingestion, staged shedding, per-SKU power capping • Land a utility-validated pilot with fast ramp, tight accuracy, and high-fidelity telemetry • Deliver dynamic power management for oversubscription • Build workload-aware GPU power estimation validated against fleet telemetry • Design for safety: authenticated signal ingress, bounded blast radius, fail-safe defaults • Own graceful ride-through of power-loss and grid-stress events • Partner with Data Center Engineering and Energy teams on interconnection and integration • Integrate with cloud control plane across Kubernetes and Slurm fleets • Hire and lead the team from the ground up **Required Experience:** • 12+ years in software engineering; 5+ years leading engineering teams • Deep experience with distributed control planes, orchestration, or fleet automation • Track record with safety-critical or physically-actuating systems • Working knowledge of data-center power systems and GPU power management • Experience with energy markets or grid programs (demand response, curtailable-load tariffs, ISO/RTO signals) • Strong build-vs-buy judgment and hiring acumen **Bonus:** GPU cluster operations, energy sector background, GPU power/performance modeling, checkpoint/preemption experience **Compensation:** $285,000 - $335,000 + Bonus + RSUs **Benefits:** Competitive equity, comprehensive health/dental/vision, HSA contributions, parental leave, professional development, wellness support, 401(k) match, and more.
Listing freshness
CronJobs last confirmed this listing 17h ago. If its source stops confirming the opening for seven days, this page is removed from active inventory.