Senior RDMA & AI Networking Performance Engineer — Remote

Codilime; Warszawa; Remote; PLN 120,000 - 170,000; Full time

Benefits offered by this job

Flexible work arrangements

Training budget

Job summary

CodiLime is looking for an RDMA Performance Engineer to join the project team working on SONiC validation on NVIDIA Spectrum-Based Platforms for AI Networking. You will configure, troubleshoot and optimize data center fabrics to achieve maximum throughput and minimal latency.

You should have extensive experience with network engineering, RDMA, NCCL, RoCEv2, PFC and ECN, and a hands-on track record with NVIDIA/Mellanox Spectrum switches and ConnectX NICs.

Qualifications

• 7+ years of professional experience in network engineering, with preferably 3+ years specifically focused on data center network design and architecture.
• Hands-on experience configuring and troubleshooting core network protocols and technologies, including BGP, EVPN/VXLAN, LACP, and ECMP.
• Deep understanding of AI-dedicated data center designs, including Rail-Optimized Design (ROD) and Rail-Unified Design (RUD) topologies.
• Proven experience configuring, managing, and troubleshooting NVIDIA/Mellanox Spectrum switches and ConnectX Network Interface Cards (NICs).
• Deep understanding of the protocols and congestion control mechanisms required for AI-dedicated data centers, specifically RoCEv2, Priority-based Flow Control (PFC), and Explicit Congestion Notification (ECN).
• Demonstrated ability to tune and optimize network environments, with specific focus on switch buffer tuning and the implementation of Quality of Service (QoS) policies.
• Experience conducting performance benchmarking and validation for AI-ready data center fabrics utilizing RDMA and the NVIDIA Collective Communication Library (NCCL).
• Demonstrated ability to leverage AI agents/tools to automate tasks, testing, and analysis
• Strong technical documentation skills and professional English communication
• Nice-to-have certifications (CCNP/CCIE Enterprise/DC, JNCIP) and exposure to CI/CD and IaC concepts

Responsibilities

• Configure, troubleshoot, and maintain NVIDIA/Mellanox Spectrum switches and ConnectX NICs.
• Implement and tune congestion control for AI workloads: RoCEv2, PFC, ECN.
• Optimize networks for maximum throughput, low latency, and minimal JCT via setup tuning.
• Perform performance benchmarking and validation on AI-ready data center fabrics using RDMA and NCCL.
• Leverage AI agents/tools to automate routine tasks, testing, and analysis.
• Author and maintain clear technical documentation detailing designs, configurations, and benchmarks.

Skills

Network design

BGP

EVPN/VXLAN

LACP/ECMP

RoCEv2

PFC

ECN

RDMA

NCCL

QoS

English communication

Tools

pyATS

NAPALM

Batfish

Nornir

NetBox

GitLab CI

GitHub Actions

Jenkins

Job description

CodiLime is looking for an RDMA Performance Engineer to join the project team working on SONiC validation on NVIDIA Spectrum-Based Platforms for AI Networking. You will configure, troubleshoot and optimize data center fabrics to achieve maximum throughput and minimal latency.

You should have extensive experience with network engineering, RDMA, NCCL, RoCEv2, PFC and ECN, and a hands-on track record with NVIDIA/Mellanox Spectrum switches and ConnectX NICs.