AI Compute GPU SME — Fault Isolation & RCA Lead

Kyndryl; Jakarta Pusat; On-site; IDR 300,000,000 - 600,000,000; Full time

Job summary

NVIDIA is seeking a senior technical authority for GPU compute infrastructure, responsible for leading fault isolation across GPUs, CPUs, memory and interconnects. You will own deep diagnostic analysis, establish health baselines and define escalation criteria for L1/L2 teams, ensuring robust firmware/BMC strategies and risk assessments.

You will drive RCA investigations, coordinate OEM recommendations, and mentor junior engineers while maintaining safe service boundaries on liquid-cooled

Qualifications

• 5-8+ years of advanced server/HPC/GPU infrastructure engineering with production support.
• Deep experience with NVIDIA datacenter GPUs and large multi-GPU systems (GB200 NVL72/B300 desirable).
• Expert-level BMC/Redfish/IPMI, PCIe, GPU health and hardware diagnostics; strong DCGM/nvidia-smi and error-analysis capability.
• Strong understanding of NVLink/NVSwitch topology, GPU interconnect failure modes and driver/firmware relationships.
• Strong Linux diagnostic skills and ability to distinguish hardware, firmware, driver and workload symptoms.
• Demonstrated ownership of complex RCA, major incidents, lifecycle changes and OEM engineering escalations.
• Experience with liquid-cooled high-density GPU platforms and safe service boundaries preferred.
• Must demonstrate independent technical resolution capability. Candidate experience not limited to ticket dispatch or vendor coordination.

Responsibilities

• Lead advanced fault isolation across GPU, CPU, memory, motherboard, PCIe, NVLink/NVSwitch, BMC, firmware, power and cooling-related symptoms.
• Perform deep analysis of XID/ECC events, device enumeration, PCIe errors, NVLink/NVSwitch state, performance degradation and recurring hardware instability.
• Define diagnostic decision trees, health baselines, acceptance criteria and escalation thresholds for L1/L2 teams.
• Own firmware/BIOS/BMC compatibility strategy, maintenance validation, rollback criteria and complex change risk assessment.
• Lead technical RCA/problem-management investigations and identify permanent corrective actions rather than ticket-level workarounds.
• Direct advanced FRU/tray replacement and post-maintenance validation within OEMauthorized boundaries, including liquid-cooled platform considerations.
• Correlate compute symptoms with adjacent fabric, storage, OS/driver and cooling domains and bring in the correct SME when required.
• Challenge and coordinate OEM recommendations, provide evidence-quality escalation packages and remain accountable through service restoration.
• Mentor L2 engineers, review runbooks and acceptance procedures, and act as technical authority during P0/P1 incidents and commissioning.

Skills

Advanced GPU compute infra

NVIDIA datacenter GPUs

NVLink/NVSwitch topology

Linux diagnostics

DCGM/nvidia-smi diagnostics

BMC/Redfish/IPMI

Troubleshooting large multi-GPU

Independent technical resolution

Tools

BMC/Redfish/IPMI tooling

PCIe diagnostics tools

Job description

NVIDIA is seeking a senior technical authority for GPU compute infrastructure, responsible for leading fault isolation across GPUs, CPUs, memory and interconnects. You will own deep diagnostic analysis, establish health baselines and define escalation criteria for L1/L2 teams, ensuring robust firmware/BMC strategies and risk assessments.

You will drive RCA investigations, coordinate OEM recommendations, and mentor junior engineers while maintaining safe service boundaries on liquid-cooled