About the company
Graphcore is a globally recognized leader in AI computing systems, part of the SoftBank Group, designing advanced semiconductors and data centre hardware.
Responsibilities
- Maintain all network fabrics and network topologies operating at peak performance on Graphcore Data centers housing AI server platforms
- Collaborate with the Network architecture team through all phases of the design and development lifecycle
- Primary support of deployed internal fabrics in AI Datacenters
- Direct escalation for Hardware, Link and Performance issues on network fabrics
- Escalation support for customer network fabrics
- Triage day-to-day network topologies (scale-up, scale-out, and front-end networks) issues
Requirements
- Bachelor's, master's degree or equivalent experience in Network Engineering/Security, IT, Computer Science, Computer Engineering, or related field
- 9+ years of hands-on experience in network engineering and 3+ years supporting networking on AI or Hyperscale Datacenters
- Knowledge of RDMA and its implementations, specifically ROCEv2 or InfiniBand
- Experience with congestion control and lossless Ethernet technologies including PFC, ECN and DCQCN
- Experience with 400G and 800G network fabrics and optical transceivers
- Experience sustaining non-blocking, multi-tier CLOS networks (Spine / Leaf / Super-Spine)
- Extensive troubleshooting experience for BGP and OSPF as well as VXLAN implementations
- Cross-functional collaboration with network architects, hardware engineers, AI software engineers, and Datacenter infrastructure teams
- Differentiators: NetDevOps with Terraform or Ansible, scripting in Python, Go, or JSON, hardware from Broadcom, Arista, Cisco, or Pensando, telemetry protocols such as gNMI, Sflow, Netflow, Redfish, Prometheus
Conditions
- Flexible working
- Medical, dental and vision coverage
- FSAs and HSAs
- Disability and life insurance
- 401(k) retirement plan
- Commuter benefits and wellness services
- Employee Assistance Programme