Principal Software Engineer, GPU Firmware and GPU System Software — CSP Engagements

NVIDIA•Austin, CA

About The Position

We're looking for a Principal Software Engineer to join our CSP Engagements team as the technical focal point for GPU firmware and GPU system software, working directly with engineering teams of key CSP / hyperscale customers to ensure they can reliably manage, update, and operate NVIDIA GPU firmware at fleet scale. You will drive work streams with engineering teams of key CSPs/hyperscale customers to build shared understanding of GPU firmware and system software integration, incorporate their feedback into NVIDIA's feature roadmap and delivery plan, and ensure customer-side automation and recovery procedures are ready before each firmware release. Your cross-CSP visibility enables you to identify patterns in GPU firmware operational challenges that drive systemic improvements no single customer engagement could surface alone.

Requirements

15+ years of experience in GPU system software, GPU firmware, or accelerator platform engineering.
BS or MS in Computer Science, Electrical Engineering, or related field (or equivalent experience)
Deep understanding of GPU architecture internals: streaming multiprocessors, GEMM execution, compute kernels, memory hierarchy, and how firmware/driver decisions impact GPU compute performance
Understanding of multi-GPU fabric architectures (NVLink, or similar) and how firmware coordinates across multiple GPUs in a rack-scale system
Understanding of GPU firmware architecture: VBIOS, GPU microcontroller firmware, InfoROM, and their interaction with the GPU driver stack
Experience with firmware update lifecycle management at scale: multi-device update sequencing, A/B updates, rollback, staged rollout, emergency recovery
Understanding of GPU error handling and recovery flows — how firmware-level errors propagate through the driver stack to application-visible failures
Experience with GPU health monitoring and telemetry: Xid errors, thermal events, power events, ECC counters, and their significance for firmware/software teams
Customer obsession — genuine passion for simplifying GPU firmware integration for fleet-scale customers.
Proven success influencing engineering teams to improve quality and fleet manageability

Nice To Haves

Direct experience with NVIDIA GPU VBIOS, GPU microcontroller firmware, or GPU driver internals
Background in GPU fleet management at 10K+ GPU scale — firmware rollout, health-based remediation, fleet-wide configuration management
Experience with GPU error taxonomy (Xid classification, NVLink error counters, ECC events) and building runbooks around GPU firmware behavior
Understanding of GPU security: secure boot chain, code signing, attestation, debug authentication, multi-tenancy isolation at the firmware level
Familiarity with GPU power management architecture and its impact on workload performance at fleet scale

Responsibilities

Drive GPU firmware & siftware work streams with CSP engineering teams — ensuring they understand GPU firmware architecture (VBIOS, InfoROM, microcontroller firmware), update sequencing, recovery procedures, and GPU power management
Gather and synthesize CSP feedback on GPU firmware/software — covering manageability, observability, security requirements (e.g., multi-tenancy isolation, secure boot, attestation), and performance — and champion those priorities into NVIDIA's GPU firmware/software feature roadmap and delivery plan
Drive GPU firmware update orchestration for large-scale deployments — multi-GPU update sequencing, rollback strategy, failure handling, and validation across hundreds of GPUs per rack
Serve as the technical focal point between NVIDIA and CSP firmware/software engineering — ensuring GPU behaviors (error recovery flows, thermal protection, power state transitions) are well-documented and accessible for customer integration
Identify cross-CSP GPU SW/FW issue patterns — common update failures, recovery gaps, and configuration problems — and drive documentation, tooling, and test strategy improvements