Senior Network & Interconnect Engineer

We are looking for a

Senior Network & Interconnect Engineer.

The successful candidate will design, deploy, and operate production-grade

InfiniBand

and

AI-optimized Ethernet

fabrics, ensuring scalability, performance, and reliability for research institutions and enterprise customers across Italy and Europe. The role involves designing and operating Spine-Leaf datacenter fabrics (InfiniBand and Ethernet with BGP/EVPN/VXLAN), leading technical escalations with NVIDIA Networking, and evaluating next-generation interconnect technologies such as DPUs, CXL fabrics, photonic networking, and Ultra Ethernet.

Potrebbe essere proprio lei il candidato/la candidata ideale per questo lavoro. Legga tutte le informazioni associate e invii la sua candidatura.

Key Responsibilities: Design datacenter fabric and HPC interconnect end-to-end: topology selection (fat-tree, dragonfly, rail-optimized), oversubscription, bisection bandwidth targets, failure-domain design, and integration with storage backends (Ceph, VAST, WEKA, BeeGFS over IB / RoCE). Bring up, tune, troubleshoot, and operate

InfiniBand XDR / NDR

(Quantum-3 / Quantum-2) and

Ethernet

AI-optimized fabrics: subnet manager (OpenSM) configuration, partitioning, QoS, adaptive routing, congestion control, and performance counters. Design and operate Spine-Leaf datacenter fabrics with

BGP / EVPN / VXLAN , multi-fabric and multi-site interconnect, across switch ecosystems including Mellanox SN, SONiC, Dell OS10 etc. Own the physical layer of AI Factory and HPC cluster networking: optical cabling design (BiDi, parallel optics, DAC, AOC), OOB management, patching matrix, and power and space planning in collaboration with the Operations and Facilities teams. Instrument fabric observability: sFlow, IPFIX, streaming telemetry (gNMI), Prometheus exporters for switch fabric; correlate with the team’s monitoring stack and drive capacity planning. Contribute to R&D and presales initiatives: author the networking sections of EuroHPC JU and EU bids; participate in technology scouting on DPU (BlueField-3+), CXL fabric, photonic interconnect, and Ultra Ethernet Consortium developments. Write and maintain

runbooks

— step-by-step operational procedures with explicit go / no-go gates and rollback paths — for fabric bring-up, upgrade, and incident response workflows. Mentor junior and mid-level engineers in the team on advanced networking and fabric operations.

Education: Master’s degree or Ph.D. in Telecommunications, Computer Networks, Computer Science, or a related STEM field — or equivalent practical experience, including: 4+ years

of hands-on engineering experience in datacenter networking and high-speed interconnect. 2+ years

of direct experience operating InfiniBand fabrics in production (subnet management, troubleshooting, performance tuning). 2+ years

of hands-on Spine-Leaf datacenter fabric design and operations at scale (BGP / EVPN).

SKILLS AND COMPETENCES

Core Technical Skills: InfiniBand:

Mellanox HCA (ConnectX-6 / 7 / 8, BlueField), Quantum-2 / 3 switches, OpenSM configuration and tuning, IB performance counters, IB-to-Ethernet gateway. Ethernet AI fabric:

Spectrum-X, RoCEv2 with DCQCN / HPCC, PFC tuning, adaptive routing, congestion control. Network protocols:

BGP (including EVPN), OSPF, MP-BGP, VXLAN, Geneve, MLAG, MC-LAG, RSTP, LLDP. Linux networking:

netlink, tc qdisc, eBPF for networking, XDP basics. Automation & Infrastructure as Code:

Python (production quality), Ansible with custom modules and Bash scripting. Go is a plus. Problem-Solving & Documentation:

Systematic troubleshooting methodology across hardware, firmware, driver, and OS layers. Ability to write design documents, RCA reports, and runbooks. Communication:

Fluent English (working language with Mellanox / NVIDIA / Dell engineering); clear written and verbal communication for vendor, customer, and internal audiences.

Nice-to-Have Skills: Fabric management & orchestration platforms NVIDIA UFM

(Enterprise / Telemetry / Cyber-AI) for InfiniBand fabric management and anomaly detection. NVIDIA NetQ

for Cumulus / SONiC fabric telemetry. Dorado Cruz

(Cruz Fabric Controller / Operations Center) for multi-vendor SONiC fabric orchestration. Exposure to intent-based platforms such as Juniper Apstra, Cisco Nexus Dashboard, or Arista CloudVision.

Soft Skills: Strong technical leadership and team coordination skills. Excellent communication and stakeholder management abilities. Mentoring mindset with a passion for developing team members. Strong analytical, problem-solving, and decision-making skills. Ability to prioritize, organize, and deliver results in dynamic environments. Collaborative approach with experience working across cross-functional teams. Proactive attitude, strong sense of ownership, and commitment to continuous improvement. Adaptability and resilience in fast-paced, mission-critical environments.

Working environment: Remote work

— Flexibility to work from anywhere in Italy, with no hybrid mandate; on-site presence required only about once a month, with travel covered by the company. Objective-based work

— Clear, measurable objectives reviewed and updated throughout the year, with a structured growth path and access to cutting-edge interconnect technology in production. On-site missions

— The role involves periodic on-site engagements at customer and partner facilities in Italy and across the EU (cluster bring-up, vendor PoC, co-deployment), with travel covered by the company. RAL : 50.000,00 €

Early Career Opportunities While this position targets experienced engineers,

we are also building the next generation of networking specialists . If you have a solid technical foundation, strong curiosity, and a passion for high-performance networking, we encourage you to apply even if you do not yet meet every requirement. xpavfwm Junior engineers will have the opportunity to grow through mentorship.