$100,000 - $189,750 USD yearly
Develop power estimation models and architecture exploration tools for NVIDIA DGX and multi-GPU datacenter platforms.
Originally posted September 10, 2026 by the employer. Relisted by the employer on September 11, 2026.
About the role
NVIDIA's Applied Power Architecture team is hiring an AI GPU Power Architect to contribute to energy-efficient GPU and SoC architectures. The role focuses on developing power estimation models and tools for GPU products and systems like NVIDIA DGX.
What you'll do
- Build and contribute to power estimation models and tools for GPU products and systems including NVIDIA DGX.
- Conduct early GPU and system architecture exploration focusing on energy efficiency and TCO improvements at the GPU and datacenter level.
- Perform performance vs. power analysis across NVIDIA product lineups.
- Deploy machine learning techniques to develop accurate power and performance models for GPUs, CPUs, switches, and platforms.
- Analyze GenAI and HPC workload characteristics at datacenter scale (multi-GPU) to define hardware/software features for Perf@Watt improvements.
- Model and analyze high-speed and high-density interconnect technologies.
What you'll need
- Bachelor's or Master's degree in Electrical Engineering, Computer Engineering, or equivalent experience.
- Understanding of computer architecture and digital logic design / VLSI design.
- Proficiency with Python and data analysis packages including Pandas, NumPy, and PyTorch.
- Knowledge of energy-efficient chip design fundamentals, including low-power design techniques such as multi-VT, clock gating, power gating, and dynamic voltage-frequency scaling (DVFS).
Skills: GPU architecture, power estimation, energy efficiency, DVFS, Perf/Watt, ML
More semiconductor jobs: NVIDIA jobs · SoC design jobs
Tracked in our index since 2 May 2026. This is our observation date, not the employer posting date. Specialty: SoC/Chip Architecture (SOC Architecture). 541 active roles in this specialty at 77 employers in our index, as of 21 Nov 2026.
How we counted: postings read from employer career sites. How we count