Senior System Software Engineer - GPU Power Management

NVIDIA•Austin, CA

About The Position

NVIDIA is redefining accelerated computing and powering the next era of artificial intelligence. Our data-center platforms bring together GPUs, CPUs, networking, systems, and software to solve some of the world’s most exciting computing problems. We are looking for a Senior System Software Engineer to join NVIDIA’s GPU Performance and Power Management Software team. You will be responsible for architecting production software for performance states (P-states) and Power Management Controllers. Your work will also cover dynamic voltage and frequency scaling (DVFS) and power-management control system technologies for next-generation data-center GPUs. You will work at the intersection of hardware architecture, embedded firmware, device drivers, operating systems, and AI workloads. Your responsibilities will span the entire product lifecycle—from initial architectural design and pre-silicon validation through silicon bring-up, productization, and deployment in extensive data-center infrastructures. This is an outstanding chance to influence how next-generation NVIDIA GPUs provide top-tier performance, energy efficiency, and reliability.

Requirements

BS or MS degree in Computer Science, Computer Engineering, Electrical Engineering, or a related field—or equivalent experience.
At least 12+ years of extensive industry experience crafting system software, embedded firmware, device drivers, or other hardware-related production software.
Strong programming skills in C, including experience developing, debugging, and maintaining complex low-level software.
Deep understanding of operating-system fundamentals, device-driver architecture, embedded or real-time software, concurrency, interrupt handling, and hardware programming.
Experience interpreting hardware specifications and developing software interfaces for registers, telemetry, interrupts, and control mechanisms.
Strong understanding of computer architecture and hardware-software interactions.
Proven ability to independently debug complex problems across multiple software and hardware layers.
Excellent written and verbal communication skills, with a track record of working effectively across organizational and geographic boundaries.

Nice To Haves

Hands-on experience with GPU, CPU, accelerator, or SoC power and performance management.
Background with crafting or implementing P-state selection, DVFS, clock-control, voltage-control, power-capping, thermal-management, throttling, or workload-aware control policies.
Experience developing real-time feedback controllers, including reasoning about stability, responsiveness, hysteresis, latency, and noisy telemetry.
Understanding of electrical and platform considerations such as voltage-frequency relationships, transient power, thermal limits, power delivery, and system-level power budgets.

Responsibilities

Architect, build, implement, and debug GPU power- and performance-management software, with particular emphasis on P-state management and DVFS.
Develop real-time controllers, policies, and mechanisms that dynamically optimize GPU performance, power, thermals, and energy efficiency under demanding data-center workloads.
Drive features across the full product lifecycle: requirements, architecture, implementation, pre-silicon validation, silicon bring-up, productization, and production support.
Collaborate with GPU architects and hardware designers to define hardware-software interfaces and influence the power-management capabilities of next-generation processors.
Analyze complex interactions among workloads, clocks, voltages, power limits, thermals, telemetry, and system-level policies.
Lead the investigation and resolution of complex power, performance, stability, and reliability issues spanning firmware, drivers, hardware, and platform software.
Develop validation strategies and automation for functional correctness, transition latency, performance-per-watt, and robustness across operating conditions.
Build clear architecture, interface, and compose specifications for new power-management features.
Partner with geographically distributed architecture, ASIC, firmware, driver, validation, platform, and data-center systems teams.