The opportunity
Scale GP (Scale Generative AI Platform) is Scale’s enterprise AI platform, providing APIs and infrastructure for knowledge retrieval, inference, evaluation, agents, and more. We deploy SGP across AWS, Azure, and GCP, often directly into customer-controlled cloud environments…
What you'll do
Own network architecture for the SGP platform: define and enforce network standards across cloud environments (AWS, Azure, GCP) and customer deployments
Design and review VPC architectures, peering, DNS, load balancing, and: CDN/edge configurations (e.g., Cloudflare), grounding root-cause and tradeoff discussions in data and domain depth
Bring strong technical judgment to VPN, routing, access, ingress/egress, and traffic-flow decisions
Review proposed networking solutions from platform, product, and: forward-deployed teams; evaluate what should and should not be introduced into the platform boundary
Define the standard connectivity pattern between Scale's control plane and: customer data planes (VPC peering, PrivateLink/Private Service Connect, site-to-site VPN, reverse tunnels) — and converge today's per-customer designs onto it
Own the customer-boundary delivery blueprint: how code, images, and traffic cross into a customer tenant — CI/CD mirroring across org boundaries (including customer-side Azure DevOps), artifact scan gates, private registries, ingress — so new engagements configure a pattern instead of designing one
What they're looking for
- + years of network engineering experience, including designing and operating: production networks in cloud environments
- Deep expertise in cloud networking on at least two of AWS, Azure, and GCP:: VPC design, peering, Transit Gateway/hub-and-spoke topologies, private connectivity (PrivateLink, Private Service Connect), and DNS
- Kubernetes networking depth: CNI, ingress, NetworkPolicy, service mesh (Istio/Envoy) — this is where most of our real incidents live
- Strong fundamentals in TCP/IP, BGP, routing, firewalls, VPN (site-to-site and client), and TLS
- Experience with edge/CDN and traffic-management platforms such as Cloudflare
- Has defined a network standard or reference architecture that other teams: adopted, and enforced it through design review
- Comfortable as the sole domain owner: able to collect requirements across five-plus live environments you didn't design, then converge them without breaking any
- Has made build-vs-adopt networking calls with vendor-support or contractual consequences in a customer's cloud
