Lead System Software Architect, AI/GPU Platforms

AMD; Austin (TX); On-site; USD 180,000 - 240,000; Full time

Job summary

AMD is seeking a Systems Software Architect to join the MI400-class subsystems team, focusing on next‑generation, rack‑scale AI platforms. You will define software stack interfaces across kernel drivers, ROCm runtime, and firmware, collaborating with silicon teams to deliver scalable performance.

The role emphasizes reliability and performance at scale, with experience in GPU compute ecosystems and collaboration with AI workloads. Austin, TX and Santa Clara, CA locations apply.

Qualifications

• Linux memory management and heterogeneous memory management knowledge.
• GPU/DRM driver development experience.
• Cache coherence and memory consistency knowledge.
• GPU networking technologies incl. NVLink, UALink, RDMA.
• Scale-up/scale-out networking and RCCL/NCCL experience.
• Familiarity with data-center fabrics like Infinity Fabric/InfiniBand/RoCE.

Responsibilities

• Own end-to-end system software architecture for MI400-class subsystems (GPU memory management, scheduling, RAS).
• Drive architecture across kernel driver, user-runtime, firmware interfaces, ROCm platform.
• Collaborate with silicon/SoC architects during pre-silicon definition to shape interfaces and contracts.
• Define software strategy for multi-GPU and rack-scale topologies and interconnects.
• Establish reliability, availability, serviceability, telemetry, and graceful degradation at scale.
• Identify bottlenecks in launch path, memory subsystem, and communication paths and propose software solutions.
• Produce architecture specifications, reference designs, and design reviews aligning firmware, driver, runtime, and frameworks.
• Engage with AMD customers and translate workload requirements into architectural direction.

Skills

Linux memory mgmt

GPU driver

Cache coherence

GPU networking

Scale-out networking

Infinity Fabric

Education

Bachelors/Masters in EE/CE/CS

Job description

AMD is seeking a Systems Software Architect to join the MI400-class subsystems team, focusing on next‑generation, rack‑scale AI platforms. You will define software stack interfaces across kernel drivers, ROCm runtime, and firmware, collaborating with silicon teams to deliver scalable performance.

The role emphasizes reliability and performance at scale, with experience in GPU compute ecosystems and collaboration with AI workloads. Austin, TX and Santa Clara, CA locations apply.