Advanced Power Methodology Engineer


Location Pin Icon
Santa Clara, CA

Job Description

As posted by the hiring company

Job Overview:

We are now looking for an Advanced Power Methodology Engineer!

At NVIDIA, we pride ourselves on having energy-efficient products. We believe that continuing to maintain our products' energy efficiency compared to the competition is key to our continued success. Our team is responsible for researching, developing, and deploying methodologies to help NVIDIA's products become more energy efficient.

You will invent novel techniques to model, analyze, and reduce power consumption. You will define research topics, plan and implement them, to gain new and key insights on how and where power is consumed. You will develop methodologies to apply knowledge from your research to make our products more efficient. As a senior member of the Power Modeling & Methodology Team, you will collaborate with Architects, Software Engineers, ASIC Design Engineers, Performance and Physical Design teams to study and implement low power methodologies for NVIDIA's next-generation GPUs and Tegra SOCs. Your contributions will help us improve idle power and active power for various workloads, by influencing architectural, design, and power management improvements.

What you'll be doing:

  • Research, develop and own advanced AI/ML/DL methodologies to estimate pre-silicon power and improve GPU energy efficiency.

  • Develop and own methodologies and workflows to train power models using ML and/or statistical techniques.

  • Build efficient data pipelines to gather power data from different sources, such as simulation, emulation, for developing advanced data-dependent methodologies.

  • Develop tools and infrastructure to mine data from pre-silicon and post-silicon performance runs to identify GPU utilization, power consumption, and energy usage.

  • Define and implement tools and methodologies for efficient integration of power models with performance tools.

  • Develop tools to debug energy inefficiencies observed in various workloads run on silicon, RTL, and architectural simulators. Identify and suggest solutions to fix the energy inefficiencies.

  • Work with power model developers to identify bottlenecks in the flow and speed up model development and integration.

  • Experiment with various ML techniques to improve the accuracy of power models.

  • Enable efficient storage and retrieval of data from databases.

  • Enable easy visualization of data using platforms such as Kibana.

What we need to see:

  • MS (or equivalent experience) with proven experience or PhD in related fields.

  • 2+ years of experience.

  • Strong coding skills, preferably in Python, C++.

  • Ability to formulate and analyze algorithms, and comment on their runtime and memory complexities.

  • Background in machine learning, AI, and/or statistical modeling.

  • Basic understanding of VLSI, digital design, and computer architecture concepts.

  • Exposure to fundamental concepts of power and energy consumption, estimation, and low power design.

  • Desire to bring quantitative decision-making and analytics to improve the energy efficiency of our products.

  • Good verbal/written English and interpersonal skills.

NVIDIA is committed to fostering a diverse work environment and proud to be an equal opportunity employer. As we highly value diversity in our current and future employees, we do not discriminate (including in our hiring and promotion practices) on the basis of race, religion, color, national origin, gender, gender expression , sexual orientation, age, marital status, veteran status, disability status or any other characteristic protected by law.