Power Architect for AI Accelerators

MatX; Mountain View (CA); Hybrid; USD 120,000 - 600,000; Full time

Benefits offered by this job

PTO & holidays

Remote work up to 3 weeks

Health insurance & dental/vision

Life insurance & disability

401K/retirement match

Professional development budget

Commuting benefits

AI resources budget

Job summary

MatX is seeking a Power Architect to drive cross-stack power tradeoffs for next‑generation server‑class AI accelerators in Mountain View. You will define budgets, models, and tradeoffs across architecture, RTL, PD, firmware, and platform, surfacing data‑driven guidance for implementation teams.

You’ll build estimation frameworks and collaborate with architecture, RTL, DV, PD, and platform teams to optimize power, perf, area, and reliability across the product lifecycle.

Qualifications

• Must have deep experience in power architecture, modeling, methodology, or performance analysis for complex chips.
• Ability to reason across architecture, RTL, PD, firmware, platform, and thermal constraints.
• Experience building early models to influence direction and later models for correlation.
• Strong analytical skills for modeling, automation, and data analysis.

Responsibilities

• Own MatX’s shared power architecture framework for server-class AI accelerators.
• Define power budgets, models, targets, and tradeoff methods.
• Build and correlate estimation frameworks linking architecture intent to implementation.
• Partner with architecture, RTL, DV, PD, and platform teams.
• Surface and track cross-functional power tradeoffs across the product lifecycle.
• Drive decisions across power, performance, area, timing, voltage, thermal, reliability, schedule, deployability.
• Translate power assumptions and data into actionable guidance for implementation teams.
• Help define power-control features, throttling, and telemetry.

Skills

Power architecture

Power modeling

Power methodology

Power-performance analysis

Cross-stack reasoning

Analytical skills

Tradeoff communication

Job description

MatX is seeking a Power Architect to drive cross-stack power tradeoffs for next‑generation server‑class AI accelerators in Mountain View. You will define budgets, models, and tradeoffs across architecture, RTL, PD, firmware, and platform, surfacing data‑driven guidance for implementation teams.

You’ll build estimation frameworks and collaborate with architecture, RTL, DV, PD, and platform teams to optimize power, perf, area, and reliability across the product lifecycle.