Network Engineer, Design & Engineering

About this role

Design and engineer large-scale network infrastructure for AI training clusters at Fluidstack, a company building civilization-scale compute capacity. You'll own the complete network design lifecycle from customer requirements through deployment, creating lossless Ethernet fabrics and architectures for 100k+ accelerator clusters.

What you'll do

• Design end-to-end network architectures for AI training and inference workloads across multiple GPU platforms
• Create topology designs, IP addressing schemes, routing policies, and fabric configurations for front-end, back-end, and storage networks
• Develop lossless RDMA (RoCEv2) fabrics with PFC, ECN, and congestion management tuning
• Translate logical designs into physical reality including rack layouts, cabling, power constraints, and airflow considerations
• Produce high-level and low-level design documents, BOMs, cabling matrices, and lead design reviews
• Collaborate across hardware, operations, software, and validation teams to ensure designs are buildable and operationally sound

What they're looking for

• Data center network fabric design and deployment experience
• Deep expertise in CLOS/fat-tree topologies, BGP, EVPN/VXLAN
• Lossless RDMA and RoCEv2 fabric design with congestion management
• First-principles problem-solving and novel architecture reasoning
• L1-L3 networking fundamentals and tradeoff analysis
• Multi-vendor GPU platform integration
• Large-scale cluster architecture (100k+ accelerators preferred)
• Source-of-truth-driven design generation and automation

Benefits

• Competitive salary and equity compensation
• Retirement or pension plan aligned with local standards