RoCE / Infiniband Network Engineer

Job Summary

We are seeking an experienced Network Fabric Architect to design and optimize high-performance, lossless Ethernet fabrics supporting large-scale AI, GPU, and HPC environments. The ideal candidate will be responsible for architecture, deployment, and performance tuning of data center networks with a strong focus on RoCEv2/RDMA, congestion management, and scalable leaf-spine fabrics.

Key Responsibilities

• Design and architect lossless Ethernet fabrics for AI/GPU clusters and HPC workloads.
• Define network topology, segmentation, redundancy, and high-availability strategies.
• Implement and optimize DCB, PFC, ECN, QoS, BGP, and EVPN technologies.
• Lead acceptance testing, performance validation, and network readiness assessments.
• Develop telemetry, monitoring, and analytics solutions for proactive network operations.
• Collaborate with infrastructure, storage, compute, and AI platform teams to ensure optimal end-to-end performance.
• Drive network automation initiatives to improve scalability and operational efficiency.
• Troubleshoot complex network, RDMA, and application performance issues across distributed environments.

Required Skills & Experience

• 8+ years of experience in Data Center Networking and network architecture.
• Hands-on expertise with RoCEv2, RDMA, and high-performance AI networking environments.
• Strong experience designing and supporting Leaf-Spine architectures and large-scale Ethernet fabrics.
• In-depth knowledge of BGP, EVPN, QoS, DCB, PFC, ECN, and congestion control mechanisms.
• Experience with NVIDIA Spectrum-X, InfiniBand, or equivalent AI networking platforms.
• Strong understanding of optical networking, data center interconnects, and network telemetry.
• Experience with network automation tools and Infrastructure-as-Code methodologies.
• Proven ability to perform end-to-end performance analysis and troubleshooting in GPU/HPC environments.

Preferred Skills

• Experience supporting AI/ML, Generative AI, or large-scale HPC deployments.
• Familiarity with Kubernetes, cloud networking, and hybrid infrastructure environments.
• Relevant certifications such as CCNP Data Center, CCIE Data Center, NVIDIA Networking, or equivalent.