Lead AI/GPU Systems Architect

Advanced Micro Devices, Inc.; Austin (TX); On-site; USD 190,000 - 270,000; Full time

Job summary

Advanced Micro Devices, Inc. in Austin and Santa Clara seeks a Systems Software Architect to shape the software stack for MI400-class accelerators, spanning GPUs, nodes, and interconnects.

You will define interfaces, ensure performance and reliability, and partner with silicon teams to pre-silicon definitions.

This role requires deep expertise in Linux memory management, GPU drivers, and large-scale distributed systems, and offers the chance to influence world-scale AI and HPC deployments.

Qualifications

• Experience in Linux memory management with MMU/HMM.
• GPU/DRM driver development experience.
• Knowledge of cache coherence and memory consistency protocols.
• Experience with GPU networking technologies including NVLink/UALink/RDMA and peer-direct.
• Familiarity with scale-out networking and communication collectives (RCCL/NCCL, MPI, SHMEM).
• Exposure to data-center fabrics such as Infinity Fabric, UALink, InfiniBand, and RoCE.

Responsibilities

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

Skills

Linux Memory Management

DRM driver development

Cache coherence

GPU Networking

Scale-out networking

Infinity Fabric knowledge

ROCm stack familiarity

Kernels/driver/runtime knowledge

Education

Bachelor’s or Master’s in Electrical Engineering, Computer Engineering, or Computer Science

Tools

ROCm

CUDA/HIP

GPU compute ecosystems

Job description

Advanced Micro Devices, Inc. in Austin and Santa Clara seeks a Systems Software Architect to shape the software stack for MI400-class accelerators, spanning GPUs, nodes, and interconnects.

You will define interfaces, ensure performance and reliability, and partner with silicon teams to pre-silicon definitions.

This role requires deep expertise in Linux memory management, GPU drivers, and large-scale distributed systems, and offers the chance to influence world-scale AI and HPC deployments.