Drimva

NVIDIA

Senior Power Analysis and Optimization Engineer

$136K–$264K

US, CA, Santa Clara +1 · Senior · Full-time · Posted May 22, 2026

Location
US, CA, Santa Clara
US, TX, Austin
Experience
MS (or equivalent experience) and 5 years of experience, or PhD and 3 years of experience.
Education
MS (or equivalent experience) or PhD in EE/CE/CS or related fields.

About the job

Analyzes and optimizes power in GPU and SoC designs by combining power architecture expertise, automation, machine learning, and LLMs.

What you'll do

  • Analyze chip- and unit-level power data, identify root causes, and drive design improvements with cross-functional teams.
  • Develop and deploy power-aware models, flows, and LLMs for anomaly detection, analysis, and optimization.
  • Prototype architectural features in Verilog and automate power-analysis pipelines using programming languages and tools.

What they ask for

  • Strong understanding of energy consumption, power estimation, data movement, and low-power design.
  • Familiarity with Verilog and ASIC design principles, and hands-on experience with power-analysis tools or similar.
  • Solid coding and automation skills, preferably in Python, Perl, and C++.
  • Excellent communication and collaboration skills.

Nice to have

  • Experience or strong interest in machine learning, reinforcement learning, and data analytics, ideally applied to EDA, architecture, or system-level optimization.
  • Interest or experience building or using LLMs or other foundation models for engineering workflows.

Pay

136,000 USD - 218,500 USD base salary (Level 3); 168,000 USD - 264,500 USD base salary (Level 4)

Tools

Verilog, PowerArtist, PrimePower/PrimePower RTL, RTL Architect, Python, Perl, C++

AI focus

Build and apply machine learning, reinforcement learning, data analytics, and custom LLMs to analyze power, detect anomalies, identify causes, and recommend or tune power-efficient designs and policies.

Model-written summary · the original posting is authoritative

First seen Oct 16, 2026 · Last verified Nov 4, 2026 (UTC)