GPU Power Architect (AI)

TL;DR

GPU Power Architect (AI/Hardware): Designing and developing energy-efficient hardware architectures for AI/ML workloads with an accent on GPU power models and hardware mechanisms. Focus on optimizing performance-per-watt targets, mitigating di/dt transients, and creating scalable power analysis flows.

Location: On-site in Folsom, California, USA

Salary: $122,440.00 - $232,190.00

Company

A global leader in semiconductor innovation, specifically within the Data Center Group focusing on Xeon-based and custom x86-based products for HPC and AI systems.

What you will do

• Design and develop energy-efficient hardware architectures specifically for AI/ML workloads.
• Build and validate GPU power models across representative use cases and workloads.
• Architect hardware mechanisms to control ICCmax and mitigate di/dt transients.
• Conduct detailed workload power analysis to achieve performance-per-watt (perf/W) targets.
• Develop scalable flows and methodologies for accurate and efficient power analysis.

Requirements

• M.S. or Ph.D. in Computer Engineering, Electrical Engineering, or a relevant STEM field.
• Must be able to work on-site in Folsom, California, USA.
• Experience in computer architecture and digital logic design using SystemVerilog.
• Proven track record in the development and validation of GPU power models.
• Knowledge of electrical and thermal constraints in datacenter-class GPUs.
• Proficiency in Python.

Nice to have

• Ph.D. degree.
• Ability to innovate and drive strategic advancements in GPU architecture.
• Experience collaborating across diverse engineering teams to achieve complex design goals.
• Analytical mindset for evaluating tradeoffs between performance, power, and area constraints.

Culture & Benefits

• Total compensation package consisting of competitive pay and stock bonuses.
• Comprehensive benefit programs including health, retirement, and vacation.
• Opportunity to work with top-tier engineers on cutting-edge scalable and modular GPU designs.