AI Cloud Network Delivery Engineer

About the role

Bitdeer is seeking an AI Cloud Network Delivery Engineer to lead end‑to‑end network delivery for large‑scale GPU and CPU datacenters. You will coordinate vendors, manage rack layouts, and ensure high‑performance networking meets the demanding latency and bandwidth requirements of AI model training.

Key responsibilities

• Own full‑lifecycle project management for new datacenter builds and capacity expansions, authoring deployment plans and schedules.
• Plan rack layouts for massive GPU clusters, balancing power, load‑bearing, thermal constraints, and network architecture.
• Design and implement cabling architecture for InfiniBand / RoCEv2 networks, managing fiber, AOC, and DAC routing, tray paths, and labeling.
• Configure and commission Spine/Leaf switches (400G/800G), handling BGP (eBGP/iBGP), OSPF, VxLAN EVPN, and related protocols.
• Promote network automation using ZTP and Ansible to achieve error‑free, scalable configurations.
• Develop and execute network acceptance test plans, including physical link, routing convergence, and congestion control validation.
• Collaborate with business teams on GPU cluster integration testing (NCCL Benchmark, bandwidth tests) and produce delivery acceptance reports.

Required profile

• Bachelor’s degree in Computer Science, Telecommunications, Electronics or a related field.
• 5+ years of network engineering delivery experience in large‑scale internet, cloud computing, or datacenter environments.
• Hands‑on experience with AI datacenter topologies, InfiniBand or RoCEv2 physical cabling, and equipment deployment.
• Understanding of extreme latency and bandwidth demands of large‑model training.

Required skills

• InfiniBand
• RoCEv2
• Spine/Leaf switching (400G/800G)
• BGP (eBGP, iBGP)
• OSPF
• VxLAN EVPN
• ZTP (Zero Touch Provisioning)
• Ansible
• Network automation
• High‑performance cabling design (fiber, AOC, DAC)
• Rack layout planning for GPU clusters
• Network acceptance testing
• NCCL Benchmark and bandwidth testing