Senior Architecture Energy Modeling Engineer

Details

NVIDIA's Power Modeling, Methodology and Analysis Team researches, develops, and deploys methodologies that improve the energy efficiency of NVIDIA products. The team builds energy models that integrate with architectural simulators, RTL simulation, emulation, and silicon platforms. This role focuses on developing machine-learning-based power models to analyze and reduce power consumption in NVIDIA GPUs, CPUs, and Tegra SoCs.

The engineer will collaborate with architects, ASIC design engineers, low-power engineers, performance engineers, software engineers, and physical design teams to study and implement energy modeling techniques for next-generation products. The work will provide early insight into the energy consumption of graphics and artificial intelligence workloads and support architectural, design, and power-management improvements.

Responsibilities

• Work with architects, designers, and performance engineers to develop an energy-efficient GPU.
• Identify key design features and workloads for building machine-learning-based unit power and energy models.
• Develop and own methodologies and workflows to train models using machine learning and statistical techniques.
• Improve the accuracy of trained models through different model representations, objective functions, and learning algorithms.
• Develop methodologies to accurately estimate data-movement power and energy.
• Correlate predicted energy from models built at different stages of the design cycle, bridging early estimates to silicon.
• Integrate power and energy models into performance infrastructure platforms to enable combined performance and power reporting for various workloads.
• Develop tools to debug energy inefficiencies observed in workloads running on silicon, RTL, and architectural simulators, and identify potential solutions.
• Prototype new architectural features, build energy models for those features, and analyze system impact.
• Identify, suggest, and participate in studies to improve GPU performance per watt.

Requirements

• A master's degree or equivalent experience with proven experience, or a PhD, in a related field.
• Six or more years of experience.
• Strong coding skills, preferably in Python and C++.
• Background in machine learning, artificial intelligence, and/or statistical modeling.
• Background in computer architecture and an interest in energy-efficient GPU designs.
• Familiarity with Verilog and ASIC design principles is a plus.
• Ability to formulate and analyze algorithms and discuss their runtime and memory complexities.
• Basic understanding of energy consumption, estimation, and low-power design.
• Interest in applying quantitative decision-making and analytics to improve product energy efficiency.
• Good verbal, written, and interpersonal communication skills.

Compensation and Benefits

The base salary depends on location, experience, and compensation for similar positions. The base salary range is USD 168,000–264,500 for Level 4 and USD 196,000–310,500 for Level 5. The role is also eligible for equity and benefits.