Network Engineer

Role: Cisco Network Engineer

Location: London, UK

Type: Contract (Inside IR35)

Job Description:

•Design, deploy, and manage VXLAN EVPN-based fabric networks supporting AI and high-performance computing workloads.

•Build scalable spine-leaf architectures using Cisco Nexus 9000 or Arista switches, ensuring efficient Layer 2/Layer 3 segmentation and workload mobility.

•Optimize network performance for RoCEv2, GPUDirect, and low-latency traffic flows, ensuring maximum throughput and minimal jitter for GPU-accelerated clusters.

•Configure underlay and overlay protocols including BGP, OSPF, IS-IS, and troubleshoot connectivity and control-plane issues.

•Integrate data center fabric with virtualization platforms such as VMware NSX, KVM, and Hyper-V, as well as Kubernetes-based AI infrastructure.

•Collaborate with compute, storage, and DevOps teams to ensure end-to-end performance across AI training, fine-tuning, inference, and ETL pipelines.

•Drive network automation efforts using tools such as Python, Ansible, and Terraform, enabling Infrastructure-as-Code (IaC) for repeatable, scalable deployments.

•Participate in performance tuning, benchmarking, and capacity planning for evolving AI workloads and future-proof infrastructure designs.

Required Skills:

•Proven experience designing and operating VXLAN EVPN-based data center networks.

•Strong knowledge of Cisco ACI, Nexus 9000, or Arista EOS platforms.

•Expertise in data center routing and switching, including BGP, OSPF, IS-IS, multicast, and fabric path.

•Deep understanding of AI-specific networking needs, including RoCEv2, DCB, PFC, and RDMA optimization.

•Hands-on experience with automation and orchestration tools such as Python, Ansible, and Terraform.

•Familiarity with GPU-accelerated workloads and distributed systems in AI/ML and HPC environments.

Certifications:

•CCNP or CCIE Data Center (highly preferred)

•Cisco Certified Specialist – Enterprise Core or Data Center Core

•Additional certifications in NVIDIA Networking (Cumulus/Spectrum) or Arista ACE are a plus

•Experience with hybrid and multi-cloud networking for AI clusters is desirable