Platform Engineer

About the Role:

• You will own the layer where AI work physically happens: machines, isolation boundaries, and the models running on them.
• This role anchors on two systems: the inference control plane and the agent-sandboxing platform.
• You'll write Rust, build Kubernetes controllers, and develop FastAPI endpoints to manage the infrastructure.

Key Responsibilities:

• Manage the serving tier for open-weight models: engine deployment, cold-start strategy, and per-model SLOs.
• Administer the Kubernetes layer for inference and sandbox workloads: operators, autoscaling, and GPU scheduling.
• Own the stateful data plane: hosted vector stores and graph stores with backup and failover discipline.

Qualifications:

• 5+ years shipping production systems in a systems language like Rust, Go, or C++.
• Operated Kubernetes workloads that others depended on, with experience in controllers, scheduling, and autoscaling.
• Hands-on experience deploying open-weight LLM serving infrastructure and strong ability to threat-model isolation boundaries.