Data Center Network Architect
About The Position
SPAN is seeking a Network Architect to own the design, deployment, and operation of the network connecting XFRA’s distributed compute fleet. The role covers high-performance networking within GPU clusters, north-south connectivity between compute sites and customers, and wide-area architecture connecting the fleet to cloud platforms and compute partners. The architect will translate workload requirements into practical designs balancing performance, availability, security, deployment speed, and cost. This is a hands-on technical leadership role involving equipment selection and configuration, ISP service qualification, deployment automation, production troubleshooting, and development of scalable network standards. The role works closely with compute, orchestration, security, deployment, and commercial teams. SPAN describes XFRA as a distributed data center using existing infrastructure and underutilized power capacity to meet growing AI compute demand. SPAN’s broader mission is to enable electrification and support clean energy, resilient homes, and a flexible distributed electrical grid.
Requirements
- 8+ years of relevant network engineering experience, or equivalent demonstrated expertise, including ownership of production architecture and hands-on implementation.
- Strong experience with data center networks and WAN or ISP connectivity across multiple sites.
- Deep understanding of TCP/IP, BGP, IPv4/IPv6, routing, switching, segmentation, VPNs, firewalls, QoS, and high-availability design.
- Practical understanding of fiber access and carrier service models, including validating real performance, availability, and commercial service commitments.
- Experience designing and troubleshooting high-speed server networks, including NICs, optics, cabling, MTU, congestion, and oversubscription.
- Ability to translate application traffic patterns and service objectives into network requirements and clearly explain performance, resilience, and cost tradeoffs.
- Hands-on Linux troubleshooting and network automation using Python, Ansible, or comparable tools, with version-controlled configurations and disciplined change management.
- Record of taking ambiguous problems from design through deployment and production support while working effectively with engineering teams, vendors, and field operators.
Nice To Haves
- Production experience with GPU clusters, AI inference, HPC, InfiniBand, RoCE, RDMA, or NCCL performance analysis.
- Experience with EVPN/VXLAN, leaf-spine fabrics, cloud interconnects, internet peering, or operating an autonomous system.
- Experience operating a large fleet of unattended edge sites, including zero-touch provisioning, carrier-grade NAT constraints, and remote recovery.
- Familiarity with Kubernetes networking, multi-tenant compute platforms, and network-aware workload scheduling.
- Experience qualifying carrier services and negotiating technical requirements across multiple providers or regions.
Responsibilities
- Own network architecture and develop reference designs for residential nodes and larger commercial GPU clusters, including routing, switching, security gateways, addressing, management access, physical connectivity, and capacity planning.
- Design data center connectivity, including north-south paths for customer traffic, model distribution, storage access, and fleet management, as well as east-west fabrics for server communication.
- Evaluate high-speed Ethernet, InfiniBand, and RoCE against workload requirements such as bandwidth, latency, congestion, and failure behavior.
- Evaluate and qualify shared PON broadband, dedicated internet access, Ethernet private lines, wavelength services, and dark fiber where appropriate.
- Assess carrier-service performance and terms, including bandwidth, oversubscription, symmetry, latency, jitter, packet loss, addressing, routing options, service guarantees, repair commitments, lead times, and total cost.
- Work with ISPs and carriers on availability, technical requirements, provisioning, acceptance testing, and incident escalation; verify physical path diversity and shared dependencies.
- Optimize connectivity across the distributed fleet by evaluating direct internet connections, regional aggregation, private interconnects, transit, peering, cloud connectivity, traffic engineering, and failover.
- Partner with orchestration engineering to provide topology, capacity, and health signals for workload placement, traffic steering, and recovery.
- Account for model transfers, request traffic, and distributed inference communication when assessing workload efficiency within or across sites.
- Implement tenant isolation, separation of compute and management traffic, encrypted connectivity, access controls, and DDoS mitigation with the security team.
- Design remote recovery and appropriate out-of-band access for unattended sites.
- Automate repeatable provisioning, configuration validation, staged changes, rollback, and inventory management.
- Establish telemetry, service objectives, alerting, runbooks, and acceptance tests; lead network incident diagnosis and corrective actions.
Benefits
- Competitive compensation and equity grants at a well-funded, venture-backed company.
- Medical, dental, and vision base plans with 100% of employee premiums covered; options for additional coverage.
- Parental leave up to 24 weeks, depending on eligibility.
- Comfortable, sunny office space near BART and Caltrain public transit.
- Employee Resource Groups, monthly social events, SPANcakes recognition breakfast, and lunch-and-learns.
- Flexible hours, one holiday per month, and flexible time off.
Stand Out From the Crowd
Upload your resume and get instant feedback on how well it matches this job.
What This Job Offers
Job Type
Full-time
Career Level
Senior
Education Level
No Education Listed
