Senior RDMA Performance Engineer

CodiLime is a software and network engineering expert partnering with global leaders to build and maintain cutting-edge networking solutions. We are currently seeking a Senior RDMA Performance Engineer to join our team working on SONiC validation for NVIDIA Spectrum-based platforms designed for AI networking.

Key responsibilities

• Configure, troubleshoot, and maintain NVIDIA/Mellanox Spectrum switches and ConnectX Network Interface Cards.
• Implement and tune advanced congestion control mechanisms such as RoCEv2, Priority-based Flow Control, and Explicit Congestion Notification.
• Optimize network environments to ensure maximum throughput and minimal latency for AI workloads.
• Conduct extensive performance benchmarking and validation testing on AI-ready data center fabrics using RDMA and NCCL.
• Leverage AI agents and specialized tools to automate routine network tasks and analyze telemetry data.

Requirements

• 7+ years of professional experience in network engineering, with at least 3 years focused on data center design.
• Deep understanding of AI-dedicated data center topologies like Rail-Optimized and Rail-Unified designs.
• Proven expertise in configuring and troubleshooting NVIDIA/Mellanox hardware and core network protocols like BGP, EVPN/VXLAN, and ECMP.
• Strong ability to tune switch buffers and implement Quality of Service policies.
• Excellent technical documentation skills and professional English communication.

What we offer

• Professional growth supported by internal training sessions and a dedicated training budget.
• A solid onboarding process designed to ensure an easy and effective start.
• A collaborative atmosphere among passionate professionals.
• The flexibility to work on diverse and interesting projects.