Industrial AI Cloud - Network Engineer (REF5506H)

Stack

Responsibilities

• ▹Coordinate operations together with the data center, IaaS & PaaS layers: coordinate and support network lifecycle activities (installs, upgrades, changes, firmware updates) and manage network interconnections and related documentation
• ▹Switch & firewall management: provision and maintain InfiniBand switches according to ITIL standards
• ▹Automation: develop and maintain automation scripts to orchestrate the overall scope, with fine-tuning and configuration changes throughout the project lifetime
• ▹OS & firmware management: maintain network-based environments, apply patches, and manage firmware upgrades at scale
• ▹Monitoring & observability
• ▹ITIL processes: follow and improve incident, problem, and change management workflows; document runbooks and standard operating procedures; adhere to ZERO Outage guidelines
• ▹Cross-team collaboration: work closely with Platform Engineers and AI solution teams to ensure smooth deployments and operations
• ▹Manage the high-speed fabric: a unified network fabric utilizing both InfiniBand and Ethernet/RoCE technologies
• ▹Manage the management network: a separate 1 Gbps Ethernet and serial console for out-of-band (OOB) network management
• ▹PE/CE datacenter connectivity: CE routers, firewalls

Requirements

• ▹Experience in network installation, maintenance, and operations
• ▹Deep understanding of InfiniBand architecture, RDMA over Converged Ethernet (RoCE), and low-latency high-throughput networking for AI/HPC workloads
• ▹Experience with NVIDIA/Mellanox switch configuration and UFM (Unified Fabric Manager) management
• ▹Data center routing & border gateway protocols: understanding of Cisco or Juniper routers (e.g., CR-8608, PTX 10004) and BGP/OSPF routing; knowledge of ASNs, IP Transit, peering, and failover connectivity
• ▹Linux networking (Cumulus OS / Ubuntu / Debian); command-line networking skills on Linux-based systems, especially Cumulus Linux; experience configuring bridges, bonds, VLANs, and routing tables
• ▹Experience using tools such as iperf, ETHTool, nvidia-smi for network devices, perfquery, and Mellanox/NVIDIA diagnostics
• ▹Skilled in continuous monitoring, incident detection, and root-cause analysis for large-scale data center networks
• ▹Familiar with NOC/SOC operational procedures and on-call rotation models
• ▹Firewall & security management: proficient in FortiGate firewall administration — policies, NAT, VPNs, IDS/IPS, and HA configuration; understanding of security segmentation, DDoS mitigation, and zero-trust networking
• ▹Configuration & lifecycle management: hands-on experience in switch provisioning, firmware/OS upgrades, patch management, and configuration backups
• ▹Working knowledge of ITIL processes (incident, problem, change)

What we offer

• ▹Work on Europe's first industrial AI cloud with cutting-edge technologies
• ▹Direct collaboration with NVIDIA and Deutsche Telekom experts
• ▹Hybrid working model, training opportunities, and career progression

About the company

Deutsche Telekom IT Solutions, ranked as Hungary's most attractive employer in 2025 (according to Randstad's representative survey), is a subsidiary of the Deutsche Telekom Group. The company provides a wide portfolio of IT and telecommunications services with more than 5,300 employees, serving hundreds of large customers and corporations in Germany and other European countries. DT-ITS received the Best in Educational Cooperation award from HIPA in 2019 and was acknowledged as the Most Ethical Multinational Company the same year. The company continuously develops its four sites in Budapest, Debrecen, Pécs, and Szeged. NVIDIA and Deutsche Telekom are jointly developing the world's first industrial AI cloud for European manufacturers; this AI factory in Germany will host 10,000 GPUs across NVIDIA DGX B200 systems and RTX Pro Servers, with Deutsche Telekom providing secure, sovereign, and fast infrastructure including data centers, operations, security, and AI solutions. The managed network scope includes 98x MQM9700-NS2F (InfiniBand 400G), 138x SN2201, 8x SN5400, and 101x SN5610 Cumulus OS-based Ethernet switches, 4x FortiGate FG-201G and 2x FortiGate 4801F-EU firewalls, 2x border gateways (Cisco CR-8608 or Juniper PTX 10004), and 2x NVIDIA UFM appliances.