Senior Multi-GPU System Architect for AI & HPC

NVIDIA Corporation; United States; Hybrid; USD 184,000 - 357,000; Full time

Benefits offered by this job

Equity

Benefits

Job summary

NVIDIA Corporation is seeking a GPU System Architect to design multi-GPU scale-up and scale-out systems for AI and HPC datacenters. You will define architectures that tightly couple GPU compute, memory, interconnects and GPU-to-GPU fabric subsystems to deliver leading AI performance and resilience.

The role requires hands-on system-level experience, hardware-software co-design expertise, and collaboration across silicon, packaging, board, and software teams.

Qualifications

• 8+ years of system design and/or ASIC/SoC architecture for GPU/CPU or networking.
• Deep understanding of interconnects: NVLink, Ethernet, InfiniBand, CXL, PCIe.
• Experience with RDMA/RoCE or InfiniBand transport offloads.
• Proven ability to architect multi-GPU/CPU topologies with bandwidth and NUMA awareness.
• Experience with hardware-software interaction, drivers and runtimes, and distributed systems tuning.
• Strong analytical and system modeling skills (Python, SystemC, or similar).
• Excellent cross-functional collaboration with silicon, packaging, board, and software teams.

Responsibilities

• Architect multi-GPU topologies for scale-up and scale-out, balancing AI throughput and resilience.
• Define and evaluate future high-speed interconnect architectures with GPU memory systems.
• Collaborate to architect RDMA-capable hardware and optimize transport for large AI workloads.
• Use models and simulations to guide design trade-offs and bottleneck analyses.
• Work with ASIC, compiler, library and software teams for hardware-software co-design.
• Contribute to interposer, package and switch co-design for multi-die, multi-package systems.

Skills

System modeling

Python

SystemC

Hardware-software co-design

Performance tuning

Cross-functional collaboration

Education

BS in Electrical Engineering

MS in Electrical Engineering

PhD in EE/CE

Tools

NVLink

Ethernet

InfiniBand

CXL

PCIe

RDMA offloads

Job description

NVIDIA Corporation is seeking a GPU System Architect to design multi-GPU scale-up and scale-out systems for AI and HPC datacenters. You will define architectures that tightly couple GPU compute, memory, interconnects and GPU-to-GPU fabric subsystems to deliver leading AI performance and resilience.

The role requires hands-on system-level experience, hardware-software co-design expertise, and collaboration across silicon, packaging, board, and software teams.