Senior Power Architect: AI Compute & Power Optimization

Ampere; Portland (OR); On-site; USD 280,000 - 420,000; Full time

Benefits offered by this job

Health insurance

401K plan

Unlimited flextime

Paid holidays

Job summary

Ampere is a semiconductor design company focused on high-performance, energy-efficient AI compute. We seek a Distinguished Power Architect to drive power modeling, analysis, optimization, and validation across CPU, AI accelerator, memory, I/O, and interconnect designs.

This role collaborates with architecture, RTL, firmware, and silicon teams to turn workload data into design decisions. Applicants should have CMOS fundamentals and extensive experience in power analysis, RTL/gate-level methods,

Qualifications

• Experience in semiconductor power analysis and modeling.
• Strong CMOS power fundamentals and techniques.
• RTL and gate-level power-analysis experience.
• Proficiency in Python and scripting for data analysis.

Responsibilities

• Define power targets, budgets, and operating points.
• Develop and maintain power models across CPU, AI accelerator, memory, I/O.
• Run RTL- and gate-level power analysis and review results.
• Characterize workloads to identify power and performance bottlenecks.
• Drive power-management techniques including DVFS and clock gating.

Skills

Python

Tcl

Perl

C/C++

Shell

SQL

Education

Bachelor's degree in Electrical Engineering

Master's degree in Electrical Engineering

PhD in Electrical Engineering

Tools

Spyglass

Power Artist

Joules

PrimePower

Job description

Ampere is a semiconductor design company focused on high-performance, energy-efficient AI compute. We seek a Distinguished Power Architect to drive power modeling, analysis, optimization, and validation across CPU, AI accelerator, memory, I/O, and interconnect designs.

This role collaborates with architecture, RTL, firmware, and silicon teams to turn workload data into design decisions. Applicants should have CMOS fundamentals and extensive experience in power analysis, RTL/gate-level methods,