Silicon Architect: AI/ML Accelerators & HPC

Google; Sunnyvale (CA); On-site; USD 240,000 - 333,000; Full time

Benefits offered by this job

Equity grants

Bonus target

Job summary

Google LLC is seeking a Silicon Architect, AI and Infrastructure in Sunnyvale, CA. The role focuses on shaping AI/ML hardware acceleration, driving TPU architecture and integration within AI-driven systems, and leading end-to-end product architecture for next-generation accelerators.

You will collaborate with ML researchers, compiler and systems software engineers, and cross-functional silicon leads to ensure goals across functionality, performance, power, area, cost, reliability, and

Qualifications

• Bachelor's degree in Electrical Engineering, Computer Engineering, Computer Science, a related field, or equivalent practical experience.
• 15 years of experience in computer architecture.
• Experience with HPC, server processors, GPUs, or ML accelerators.

Responsibilities

• Drive end-to-end silicon product architecture and roadmap for next-generation accelerators.
• Define and own product-level chip architectures and top-level specifications.
• Lead architectural pathfinding and trade-off studies for emerging AI workloads.
• Set technical direction across subsystems to enable rapid and predictable server-class ASICs.
• Partner across silicon development phases and with compiler/runtime teams to resolve bottlenecks.

Skills

High-performance computing

Server processors

GPUs

ML accelerators

Education

Bachelor's degree in Electrical Engineering/CS/related field

15 years of architecture experience

Job description

Google LLC is seeking a Silicon Architect, AI and Infrastructure in Sunnyvale, CA. The role focuses on shaping AI/ML hardware acceleration, driving TPU architecture and integration within AI-driven systems, and leading end-to-end product architecture for next-generation accelerators.

You will collaborate with ML researchers, compiler and systems software engineers, and cross-functional silicon leads to ensure goals across functionality, performance, power, area, cost, reliability, and