NVIDIA is at the forefront of transforming computer graphics, PC gaming, and accelerated computing, now venturing into the potential of AI to redefine the next era of computing. Our GPUs are becoming the brains for computers, robots, and self-driving cars. This pioneering spirit requires vision, innovation, and top-tier talent. Join our diverse and supportive environment where inspiration drives exceptional work and lasting global impact.

This role focuses on driving comprehensive power, performance, and Perf/Watt measurements on SoC and system platforms, particularly for Windows on ARM devices.

You will characterize system behavior across various states to optimize battery life in scenarios like productivity, web browsing, and video streaming. Analyzing power telemetry, firmware traces, and OS metrics is key to identifying and resolving performance bottlenecks.

The position involves designing, building, and scaling automated power measurement infrastructure, integrating data acquisition systems, telemetry parsers, and ETL pipelines. You'll collaborate with hardware architecture, BIOS/firmware, OS, and driver teams to enable and tune new power features from pre-silicon modeling through post-silicon bring-up.

Develop advanced visualizations and detailed reports to track product health, identify regressions, and propose architectural or software optimizations to engineering and executive teams.

A strong foundation in power management concepts, including low power modes, P-states, C-states, S0ix/Modern Standby, DVFS, package power limits, and thermal throttling is essential.

Proven experience with power analysis tools on Windows/Linux platforms such as WPA, HWInfo, ETL, Xperf, and PowerCfg is required. A solid understanding of SoC architecture, including CPU, memory, I/O subsystems, and firmware (BIOS/UEFI/EC) is also necessary.

Proficiency in scripting languages like Python, PowerShell, or Bash is crucial for test automation and data parsing. The ability to analyze complex execution metrics and balance power versus performance trade-offs is a key requirement.