The GPU Cluster Health Repair Strategy team is accountable for driving repair strategy, operational execution, and cross-functional governance to keep customer GPU cluster availability at or above 97.5% at all times, where a customer cluster is defined as a shape/location combination. The team establishes KPIs, monitors performance, identifies gaps, launches corrective programs, and partners with engineering, SDE/SRE, data center operations, GSL, CHS, CPV, Warminator, TRS, customer-facing teams, and hardware partners such as NVIDIA and AMD.
This role will lead complex, high-impact program improvements at organizational scale. Defines KPIs, shapes automation, drives senior-leadership alignment, and serves as an escalation point for multi-team issues.
Internal Responsibilities
Strategic Program Leadership
Own strategic repair-transformation programs that materially improve GPU cluster availability, repair velocity, partner accountability, and operational efficiency.
Define end-to-end program strategy for high-risk availability areas, including long-running repair reduction, proactive risk detection, RMA loop improvement, spare placement optimization, healthy-node custody policy, partner hardware quality feedback, or AI-driven repair insights.
Availability and Repair Governance
Drive alignment across customer vertical TPMs, horizontal program TPMs, SDE/SRE, engineering, partner teams, and operations.
Advise leadership on SLA/SLO performance, repair productivity, operational bottlenecks, and resourcing gaps.
KPI and Data Framework Ownership
Define the KPI framework for the team, including business metrics, operational metrics, leading indicators, lagging indicators, and escalation thresholds.
Build the measurement model for shape/location health, cluster-level availability, repair aging, throughput, RMA loop time, spares sufficiency, hardware failure rate, partner blockers, and tooling efficiency.
Process and Tooling Transformation
Lead efforts to reduce manual repair execution through automation, AI-assisted triage, workflow instrumentation, and repair agent tooling.
Partner with SDE/SRE teams in India, Morocco, and Mexico to prioritize tooling investments that accelerate repair diagnosis, status visibility, escalation, and closure.
Partner and Ecosystem Influence
Lead complex engagements with partners such as NVIDIA and AMD where hardware quality, spare availability, RMA turnaround, or repair playbooks affect customer availability.
Escalation and Risk Management
Drive root-cause analysis for recurring repair failures, long-tail repair issues, spare shortages, or process breakdowns.
Make trade-offs between customer impact, business risk, technical constraints, and operational feasibility.
External Responsibilities
Strategic Program Leadership
Own strategic repair-transformation programs that materially improve GPU cluster availability, repair velocity, partner accountability, and operational efficiency.
Define end-to-end program strategy for high-risk availability areas, including long-running repair reduction, proactive risk detection, RMA loop improvement, spare placement optimization, healthy-node custody policy, partner hardware quality feedback, or AI-driven repair insights.
