Engineering Manager, Sandbox

ABOUT MISTRAL

Mistral provides full-stack AI solutions: from frontier models to developer tools, applications, and compute. We partner with enterprises tackling the hardest problems across high-stakes industries like finance, manufacturing, defense, healthcare, and the public sector, co-creating customized AI systems that they can run on their terms.

We are a dynamic, collaborative team passionate about AI and its potential to transform society. Our diverse workforce thrives in competitive environments and is committed to driving innovation. Our teams are distributed between Europe, North America, Asia and the Middle East. We are creative, low-ego and team-spirited.

THE TEAM

Mistral is building a sovereign AI cloud for Europe. Our customers in finance, healthcare, and the public sector run frontier models on European-jurisdiction infrastructure, and the Cloud Platform team builds the systems that make that possible — compute orchestration, Kubernetes, bare metal, storage, networking, and the datacenter fleet that runs it all.

We run it all on our own metal, our own network, our own datacenters — which means when something breaks, the fix is ours to write.

European Compute Units let enterprises commit to multi-year capacity on that infrastructure, and we are scaling toward a gigawatt of compute across Europe by 2030. You will design and ship the layer every sovereign AI deployment depends on and shape the team building it.

THE ROLE

Agents don't run in browsers. They run in sandboxes — and you will lead the team building ours.

You will manage the team — hiring, coaching, performance, growth paths, honest feedback — while staying close to the code: roughly half your time in architecture decisions and the hardest technical problems the team faces, the other half leading your squad.

We build secure, isolated, network-controlled environments with snapshotting primitives, designed for AI agents and reinforcement-learning workloads. An RL training run doesn't need one environment; it needs thousands, created and destroyed every second. An agent needs to be sandboxed from the world while still feeling fast enough to be useful. That tension — airtight isolation, at fleet scale, at low latency — is the product.

Your customers are the people building the next generation of AI software on our infrastructure; the primitives you ship become the runtime their agents live in.

WHAT YOU WILL DO

• Build the isolation layer: microVMs or containers, namespaces, cgroups, seccomp, snapshot and restore.
• Own the security boundary. Customers run code they didn't write and can't fully trust — your systems make that safe.
• Scale to fleets: creation latency, density, and lifecycle management for thousands of concurrent environments.
• Design and build alongside your team, owning the architecture and the hardest implementation decisions.
• Grow the engineers, not just the software — set the standard for ownership and coach people past their comfort zone.
• Own people management for the team: hiring, performance, growth paths, and honest feedback.
• Partner with Product Managers to turn frontier research needs into shippable primitives.
• WHAT WE'RE LOOKING FOR
• 3+ years of real people leadership. You have managed engineers — hired, coached, given honest feedback, and grown them. Hands-on contribution matters here, but this role is judged on the team you build and lead, not just the code you write.
• Deep Linux internals: namespaces, cgroups, seccomp, virtualization — ideally microVMs of the Firecracker class.