GFXIP Architect: Next-Gen Graphics & AI

AMD; Austin (TX); Hybrid; USD 150,000 - 210,000; Full time

Job summary

AMD is seeking a GFXIP Architecture Engineer in Austin, TX, to shape next-generation graphics and AI accelerator silicon. You will collaborate with GPU, AI, SoC, and system architects to define architectural capabilities that boost performance, reliability, and security across client, data center, and automotive markets.

You’ll design architectural mechanisms to improve diagnosability, resiliency, and recoverability while influencing production silicon through close cross-functional

Qualifications

• Bachelor's degree required; advanced degrees are a plus.
• Experience with RTL design and verification.
• Experience with GPUs, AI accelerators, SoCs, CPUs or other large-scale hardware designs.
• Familiarity with reliability metrics such as MTBF, MTBR, FIT, DPPM.

Responsibilities

• Contribute to the definition and development of architectural features for next-generation GFXIP deployments.
• Assist in writing and reviewing architectural specifications for new hardware capabilities.
• Collaborate with GPU, AI, and SoC architects on system robustness and recoverability.
• Analyze tradeoffs in performance, power, area, security, and reliability.
• Support hardware design teams in resiliency mechanisms and fault management.
• Work with cross-functional teams to ensure architectural requirements are implemented.

Skills

Architectural techniques for system R

ECC and parity concepts

RTL design and verification

GPUs, AI accelerators, SoCs

Reliability metrics (FIT, MTBF, MTBR)

Education

Bachelor's degree

Master's or PhD in relevant field is a plus

Job description

AMD is seeking a GFXIP Architecture Engineer in Austin, TX, to shape next-generation graphics and AI accelerator silicon. You will collaborate with GPU, AI, SoC, and system architects to define architectural capabilities that boost performance, reliability, and security across client, data center, and automotive markets.

You’ll design architectural mechanisms to improve diagnosability, resiliency, and recoverability while influencing production silicon through close cross-functional