Platform Engineer, Sandboxing & Air‑Gapped Infrastructure

About the role

We are looking for a platform engineer to design and operate isolated, air‑gapped environments where AI agents run safely. The platform must be installable in data‑centres without internet access and include GPU resources for high‑performance inference.

Key responsibilities

• Build single‑task, isolated sandbox environments that are torn down after each run and prepared in advance to eliminate wait times.
• Enforce outbound network policies at the network layer, keep credentials outside the sandbox, and record every action in an audit log.
• Make the entire platform—including registries, signed images, certificates and GPU drivers—installable and updatable inside an offline data‑centre.
• Provide the platform with documented guarantees of reliability and measurable performance.

Required profile

• 5+ years of platform, infrastructure or reliability engineering experience with production on‑call responsibilities.
• Deep expertise in sandboxing or execution services built on micro‑VMs or user‑space kernels (e.g., Firecracker, Kata Containers, gVisor) and offline installation pipelines.

Required skills

• Firecracker, Kata Containers, gVisor, Cloud Hypervisor
• Cilium, Talos, Zarf, Hauler
• NVIDIA GPU Operator, GPU MIG, time‑slicing
• Python, TypeScript, shell scripting
• Linux, containers, networking, Kubernetes, Helm, GitOps

What we offer

• Founding‑team scope with high impact on product direction
• AI‑augmented engineering environment
• Access to on‑premise Nvidia B200 GPUs
• Flexible remote work arrangement