Senior Network Architect – Low-Latency HPC Interconnects

Cerebras Systems; Sunnyvale (CA); On-site; USD 250,000 - 350,000; Full time

Job summary

Cerebras Systems seeks an experienced Network Systems Architect to define the scale-out and scale-up network architecture for its platforms, including proprietary interconnects and switching. You will partner with application, compiler, runtime, and systems teams to translate workloads into measurable fabric requirements and performance targets.

Your work focuses on low-latency scale-up networks with broad coverage across scale-out and customer networks.

Qualifications

• BS, MS, or PhD in Electrical/Computer Engineering or CS, or equivalent experience.
• 12+ years of industry experience, with ownership of major networking/system from architecture to deployment.
• Deep expertise in low-latency interconnects, fabric protocols, and switch architecture.
• Ability derive network requirements from incomplete workload data for trade-offs across software and hardware.

Responsibilities

• Set multi-generation architecture and roadmap for Cerebras scale-up networks.
• Collaborate with app, compiler, runtime, and libs teams to derive bandwidth and latency needs.
• Define fabric topology, protocols, and switch behavior with end-to-end considerations.
• Decide when standards-based vs custom protocols is best for system goals.
• Use models and simulations to test choices and set acceptance criteria.
• Write specs, lead reviews, and drive cross-layer decisions through implementation.

Skills

Low-latency interconnects

Fabric protocols

Switch architecture

Data-plane depth

Ethernet/IP/RDMA knowledge

Architecture decision making

Interdisciplinary coordination

Education

BS/MS/PhD in Electrical Engineering, Computer Engineering, or Computer Science

Tools

NCCL/RCCL or similar libraries

fpga/silicon bring-up tooling

Job description

Cerebras Systems seeks an experienced Network Systems Architect to define the scale-out and scale-up network architecture for its platforms, including proprietary interconnects and switching. You will partner with application, compiler, runtime, and systems teams to translate workloads into measurable fabric requirements and performance targets.

Your work focuses on low-latency scale-up networks with broad coverage across scale-out and customer networks.