Data Center Network Architect for AI Training & RoCEv2

Jobtailor; California (MO); On-site; USD 150,000 - 230,000; Full time

Job summary

Jobtailor seeks an expert Data Center Network Architect to own the design lifecycle for AI training and inference environments in the United States. You will craft topology, addressing, routing, and fabric configurations across frontend/back-end GPU clusters and storage networks, ensuring scalable, lossless performance.

Required are deep Layer 1–3 expertise, CLOS/fat-tree topology knowledge, RoCEv2 RDMA design experience, and the ability to collaborate with Hardware, DC Operations, ICT,

Qualifications

• Experience designing data center network fabrics from requirements through deployment.
• Deep Layer 1 to Layer 3 expertise required.
• Knowledge of CLOS/fat-tree topologies, BGP, EVPN/VXLAN, and underlying networking fundamentals.
• Experience designing lossless RDMA (RoCEv2) fabrics.

Responsibilities

• Own the network design lifecycle from customer requirements through deployable architectures for AI training and inference.
• Produce topology designs, IP/addressing schemes, routing policy, and fabric configuration specs across front-end, back-end GPU-to-GPU training, and storage networks.
• Adapt architectures to various accelerator platforms and form factors, rack layouts, power envelopes, and cabling approaches.
• Translate logical designs into physical infrastructure, including rack elevations, power constraints, structured cabling, fiber budgets, and airflow considerations.
• Design lossless Ethernet fabrics for RoCEv2 RDMA, including PFC, ECN tuning, traffic classes, congestion management, and distributed-training patterns.
• Produce HLDs, LLDs, cutsheets, BOMs, cabling matrices, and design decision records.
• Lead design reviews and reusable reference architectures.
• Partner with Hardware, DC Operations, ICT, Software, and Validation teams to ensure designs are buildable and operationally sound.

Skills

Data Center Network Design

Lossless RDMA (RoCEv2) Fabric Design

Layer 1 to Layer 3 Networking

CLOS/Fat-Tree Topologies

BGP and EVPN/VXLAN

Job description

Jobtailor seeks an expert Data Center Network Architect to own the design lifecycle for AI training and inference environments in the United States. You will craft topology, addressing, routing, and fabric configurations across frontend/back-end GPU clusters and storage networks, ensuring scalable, lossless performance.

Required are deep Layer 1–3 expertise, CLOS/fat-tree topology knowledge, RoCEv2 RDMA design experience, and the ability to collaborate with Hardware, DC Operations, ICT,