Staff Embedded Firmware & Accelerator Software Engineer

Google; Sunnyvale (CA); On-site; USD 207,000 - 301,000; Full time

Job summary

Google is seeking a Staff Software Engineer on the Emerging Accelerator System team to tackle low‑level software and hardware interaction. You will influence Google's accelerator roadmap and mentor across teams.

This role focuses on integrating PCIe, fabric, IOMMU, kernel and firmware work, delivering secure accelerator‑to‑host connectivity and scalable, reliable platforms for hyperscale computing.

Qualifications

• Bachelor's degree or equivalent practical experience.
• 8 years of experience in software development.
• 5 years of experience in testing and launching software products.
• 5 years of experience with embedded operating systems.
• 3 years of experience with software design and architecture.
• Experience with C, C++, operating systems, and computer architecture.

Responsibilities

• On-board emerging co-accelerators into Google's ML accelerator families, boosting performance.
• Collaborate to design and implement new system software features, including firmware and drivers.
• Design tests, telemetry, and dashboards to monitor and debug potential issues.
• Analyze and resolve complex system-level challenges in host and management software.
• Provide technical leadership to drive software plans and cross-functional dependencies.

Skills

Software development

Software testing

Embedded operating systems

Software design

C/C++

Education

Bachelor's degree or equivalent practical experience

Master’s degree or PhD in Engineering, Computer Science, or related technical field

Job description

Google is seeking a Staff Software Engineer on the Emerging Accelerator System team to tackle low‑level software and hardware interaction. You will influence Google's accelerator roadmap and mentor across teams.

This role focuses on integrating PCIe, fabric, IOMMU, kernel and firmware work, delivering secure accelerator‑to‑host connectivity and scalable, reliable platforms for hyperscale computing.