Principal System Software Architect, AI/GPU Platforms - TX, Austin, United States

Originally posted 9 June 2026 by the employer. Relisted by the employer on 19 August 2026.

About the role

You will join the system software architecture team behind AMD Instinct accelerators, focusing on rack-scale AI platforms in the MI400 class spanning the GPU, node, and scale-up/scale-out fabrics. The position defines how the software stack exposes and orchestrates hardware across silicon, firmware, drivers, runtimes, and frameworks for training and inference systems.

What you'll do

- Own the end-to-end system software architecture for MI400-class subsystems, including GPU memory management, scheduling and queuing, RAS, virtualization/partitioning (SR-IOV), or scale-up/scale-out interconnect software models. - Drive architecture across the stack, spanning the kernel-mode driver (amdgpu/KFD), user-mode runtime (ROCr/HSA), firmware interfaces, and the ROCm software platform. - Partner with silicon and SoC architects during pre-silicon definition to shape hardware/software interfaces, programming models, and register/firmware contracts before tape-out. - Define software strategy for multi-GPU and rack-scale topologies, including Infinity Fabric and UALink-style interconnects, collective communication (RCCL), memory coherence, and cross-platform address translation. - Establish architecture for reliability, availability, and serviceability (RAS) at scale, covering error detection, containment, telemetry, recovery, and graceful degradation. - Produce architecture specifications, reference designs, and design reviews that align firmware, driver, runtime, and framework teams.

What you'll need

- Bachelor’s or Master’s degree in Electrical Engineering, Computer Engineering, Computer Science, or a closely related field. - System software architecture experience spanning kernel-mode drivers, user-mode runtimes, and firmware interfaces. - Experience defining hardware/software interfaces, register/firmware contracts, or programming models during pre-silicon development. - Background in multi-GPU or rack-scale topologies, memory coherence, and interconnect fabrics.

Nice to have

- Experience with Linux Memory Management and Heterogeneous Memory Management (HMM). - GPU / DRM driver development experience. - Knowledge of cache coherence and memory consistency protocols. - Experience with GPU networking and data-center fabrics such as NVLink, UALink, RDMA, peer-direct, Infinity Fabric, Ultra Ethernet, InfiniBand, RoCE, or communication collectives (RCCL/NCCL, MPI, SHMEM). - Direct experience with the ROCm stack, AMD Instinct, CUDA, or comparable GPU compute ecosystems. - Hands-on experience developing or optimizing GPU compute kernels (HIP, CUDA, Triton, or assembly-level tuning) or frameworks like PyTorch and TensorFlow.

Skills: Linux kernel, x86 computing architecture, virtualization, device drivers, ACPI, PCIe