Apply on Employer Site

Google DeepMind · 19 hours ago

System on Chip Architect

Mountain View, CA

Full-time

Onsite

Senior Level

7+ years exp

Google DeepMind is a leader in artificial intelligence, seeking a highly motivated SoC Architect to contribute to the development of innovative silicon for machine learning acceleration. The role involves defining SoC architecture and chiplet strategies for next-generation ML accelerators, requiring deep expertise in SoC design and ML-specific hardware.

Artificial Intelligence (AI)Business DevelopmentFoundational AIMachine Learning

Growth Opportunities

Responsibilities

Define and own the System-on-Chip (SoC) architecture for next-generation ML accelerators

Lead the architecture and design of the chip top-level, managing interfaces, clocking, power, and integration of all major IP blocks

Architect specific accelerator components and chiplets

Collaborate with micro-architecture and physical design teams to ensure a feasible and optimal design, making trade-offs in performance, power, and area (PPA)

Work with systems and software teams to ensure the SoC architecture meets product requirements

Participate in the system architecture definition and evaluation

Qualification

SoC architectureChiplet integrationML-specific hardwareASIC designHardware building blocksHigh-performance architecturesFlexibilityAdaptability

Required

Bachelor's degree in Electrical Engineering, Computer Science, or equivalent practical experience

7+ years of experience in SoC architecture or micro-architecture

Proven experience in chip top-level design and integration for complex ASICs

Hands-on knowledge of hardware building blocks for ML accelerators (matrix multiply units, vector engines, attention mechanisms)

Experience in evaluating trade-offs such as speed, performance, power, and area

Preferred

Master's or Ph.D. in a related field

Experience with chiplet-based designs and high-speed die-to-die interconnects (e.g., UCIe, CXL)

Knowledge of high-performance and low-power architectures for ML acceleration

Working knowledge of transformer-based large language models

Good understanding of the full ASIC design flow (RTL, verification, synthesis, PD)

Company

Google DeepMind

Glassdoor4.5

Google DeepMind aims to research and build safe artificial intelligence system to solve intelligence and advance science and humanity. It is a sub-organization of Google.

Founded in 2010

London, England, GBR

501-1000 employees