Principal Network Architect

Sharon AI, Inc; Sydney; Hybrid; AUD 180,000 - 260,000; Full time

Job summary

Sharon AI, Inc. in Sydney is seeking a Principal Network Architect to lead the networking strategy for a multi-region AI infrastructure. You will define architecture for GPU clusters, data centres and WAN connectivity, balancing performance, cost and reliability.

The role requires 10+ years in networking, deep RoCEv2/InfiniBand expertise, BGP, and IaC. You will collaborate with AI/HPC teams across Australia and international partners to scale the platform.

Qualifications

• 10+ years of networking experience with senior architecture or leadership roles.
• Bachelor's degree in CS, Engineering or IT.
• Experience designing high‑performance networking for cloud, AI, HPC, or distributed platforms.
• Proven ability to govern large-scale data centre networking for production environments.
• Expert knowledge of large-scale IP network architecture and scaling limits.
• Strong understanding of networking for distributed AI/HPC workloads and RoCEv2/InfiniBand.
• Knowledge of BGP, peering, transit and network automation with IaC.

Responsibilities

• Define Sharon AI's long-term network strategy and reference architectures.
• Act as senior technical authority for networking across multiple teams.
• Own architecture for large-scale AI training/inference environments with GPU fabrics.
• Design scalable leaf/spine and multi-tier network fabrics with performance criteria.
• Define regional, national and international backbone and data centre interconnects.
• Establish diversity requirements for carrier pathways and multi-site services.
• Architect networks for large-scale data movement, storage replication and inter-region services.
• Define IPv4/IPv6 architecture, routing policy and traffic engineering standards.
• Drive network automation, GitOps, CI/CD and IaC capabilities.
• Collaborate with AI/HPC teams to maximize GPU utilization.

Skills

10+ years networking

Network architecture

BGP routing

InfiniBand

RoCEv2

NVIDIA Spectrum-X

Infrastructure as Code

GPU cluster networking

WAN & internet edge

Collaboration across teams

Education

Bachelor's degree in Computer Science or related

Tools

NVIDIA Spectrum-X

NVIDIA BlueField DPUs

ConnectX

GPUDirect RDMA

Job description

Sydney - Hybrid preferred | Australia Remote considered

About Sharon AI

Sharon AI is building the infrastructure powering the next generation of artificial intelligence.

Operating across AI infrastructure, high-performance compute, cloud platforms and large-scale technology environments, Sharon AI delivers scalable, secure and reliable infrastructure for demanding AI, ML and HPC workloads.

The Role

As Principal Network Architect, you'll be Sharon AI's senior technical authority for networking, defining and evolving the architecture required to scale from large-scale GPU deployments into a multi-region AI infrastructure platform.

You'll bring deep expertise across both high-performance AI/HPC and GPU cluster networking, as well as WAN, carrier, internet edge and external connectivity. You'll shape the long-term network strategy across data centres, GPU clusters, storage, backbone, peering, transit, customer connectivity and cloud networking. This is a role for someone who operates several years ahead of current infrastructure demand — identifying technical and economic constraints before they become barriers to growth, while balancing performance, reliability, cost and operational simplicity.

Key Responsibilities

• Define Sharon AI's long-term network strategy, architecture, standards and reference architectures across AI/HPC clusters, data centres, storage, WAN, backbone, internet edge and customer connectivity
• Act as the senior technical authority for networking across Infrastructure, Platform, Operations, Product, Security and customer-facing teams
• Own the architecture for large-scale AI training and inference environments, including GPU cluster fabrics supporting hundreds through to tens of thousands of GPUs
• Design and govern high-performance networking architectures using technologies including NVIDIA Spectrum-X, InfiniBand, Ethernet/RoCEv2, NVIDIA BlueField DPUs, ConnectX and GPUDirect RDMA
• Architect scalable leaf/spine, super-spine and multi-tier network fabrics, with clear performance and acceptance criteria around latency, throughput, congestion, packet loss and network goodput
• Define Sharon AI's regional, national and international backbone and data centre interconnect architecture
• Establish carrier, fibre-route and physical-path diversity requirements to support resilient multi-site and multi-region services
• Architect networks capable of supporting large-scale data movement, checkpoint transfer, storage replication and customer inter-region services
• Define and own Sharon AI's long-term IPv4 and IPv6 architecture, routing policy, prefix advertisement, aggregation and traffic engineering standards
• Establish architectural boundaries, failure domains, scaling limits and expansion triggers aligned to future compute, storage, facility and platform roadmaps
• Drive scalable network automation and platform capabilities, including source-of-truth, configuration generation, GitOps, CI/CD and Infrastructure as Code
• Work closely with AI/HPC architecture teams to maximise GPU utilisation and overall cluster productivity

Skills & Experience

• 10+ years' relevant networking experience, with substantial experience in senior architecture or technical leadership roles
• Bachelor's degree in Computer Science, Engineering, Information Technology or a related discipline
• Demonstrated experience designing high-performance networking solutions for cloud, AI, HPC or distributed computing platforms
• Proven experience designing and governing large-scale data centre networking architectures supporting mission-critical production environments
• Expert knowledge of large-scale IP network architecture, including the ability to identify control-plane, data-plane and operational scaling limits
• Strong understanding of networking requirements for distributed AI and HPC workloads and expertise in high-performance east-west network design
• Strong knowledge of RoCEv2, InfiniBand and Ethernet AI fabrics, with an understanding of NCCL and collective communications
• Expert understanding of Internet BGP architecture, transit and peering
• Strong understanding of software-defined networking, Infrastructure as Code and network automation
• Experience with very large AI/HPC or hyperscale environments is highly desirable, ideally including large accelerator clusters, large BGP fabrics, multi-region cloud networks or high-capacity internet edges
• Exposure to next-generation technologies and roadmaps such as 800G/1.6T Ethernet, next-generation InfiniBand and silicon photonics is advantageous
• Experience defining customer connectivity products for cloud platforms is advantageous

Why Join Sharon AI

• Own the network architecture underpinning a rapidly scaling AI infrastructure platform
• Shape Sharon AI's transition from large-scale GPU deployments to a multi-region AI infrastructure platform
• Work across the full networking landscape — from GPU fabrics and data centres through to national and international backbone infrastructure
• Influence major architectural decisions years ahead of infrastructure demand
• Work alongside highly technical teams across AI/HPC, Infrastructure, Platform, Operations and Product
• Build scalable, repeatable and increasingly automated network architectures that will support Sharon AI's next stage of growth
• Join a highly technical and ambitious team operating at the forefront of AI infrastructure

Our Values

Integrity | Innovation | Collaboration | Wellbeing | Inclusion