Forward Deployed Engineer, Inference Infrastructure

Role overview Deploy and operate an AI inference platform within client-controlled environments, including private cloud, Kubernetes clusters, and networks without external connectivity. You will work alongside client infrastructure teams and own the serving layer from initial setup through ongoing reliability and performance.

Responsibilities - Install the platform in client VPCs, Kubernetes clusters, or air-gapped environments, integrating with their identity and key-management systems. - Build and operate model serving, GPU scheduling and autoscaling, and routing across hosted APIs, customer-managed models, or dedicated endpoints. - Adapt deployments to client network policies, release windows, and observability systems while working with their platform teams. - Maintain reusable infrastructure definitions and deployment manifests to make installations repeatable. - Monitor and own service objectives for latency, request cost, and availability after launch.

Requirements - Production experience operating Kubernetes, including scheduling, resource limits, network policies, and node failure handling. - Hands-on experience serving machine-learning models in production and making performance tradeoffs under real workloads. - Infrastructure-as-code experience with at least one major cloud provider, plus the ability to reason about fully on-premises or air-gapped deployments. - Strong communication skills for translating between product needs and client infrastructure constraints. - Willingness to travel to client sites when an in-person deployment requires it.

Benefits and work setup - Remote role with travel to client sites as needed.