[Hiring] Software Engineer @Mistral

Role Description

You will build our serverless containers product: the way developers run AI apps, MCP servers, APIs, and managed workers on our cloud without thinking about servers. Push a container, get an endpoint, scale from zero to thousands of instances and back down — across CPU and GPU.

On a hyperscaler, the serverless layer is a product someone else built. Here you build it end to end: the scheduler that decides where a container lands, the routing path a request takes, the machinery that turns a scaled-to-zero service into a warm instance before a customer notices. All of it runs on metal, network, and datacenters we own.

That tension is the product: scale-to-zero economics with cold starts in milliseconds, autoscaling that absorbs a launch spike without a config change, GPUs scheduled onto containers the way other platforms schedule CPU. The customers betting on this layer are the developers building the next generation of AI software on European infrastructure.

Qualifications

• You have shipped production systems in Go. Not dabbled. Shipped.
• You have built container or workload platforms — orchestration, scheduling, isolation, request routing — not just deployed onto someone else's serverless platform.
• Deep Linux: namespaces, cgroups, image layers, page cache — comfortable well below the application layer.
• Multi-tenancy instincts: density, isolation, and fairness when hundreds of customers share one fleet.
• Significant experience building complex backend or infrastructure systems.
• You have partnered with product, design, and business teams to ship software that reached real users.
• A low-ego, team-first mindset. We care less about your exact domain and more about whether you are the kind of technical force other engineers want to work alongside.

Requirements

• Build the container platform end to end: image delivery, placement, request routing, scale-up, scale-to-zero — across CPU and GPU.
• Make cold starts a solved problem: snapshotting, prewarming, image-layer caching, whatever it takes for scale-from-zero to feel instant.
• Treat AI workloads as the normal case: GPU-backed containers, streaming responses, long-lived connections, MCP servers that cost almost nothing while idle.
• Write Go and own the architecture and the hardest implementation decisions.
• Partner with Product Managers to turn platform decisions into a product developers want to deploy on.

Benefits

• Comprehensive benefits package designed to support your well-being, growth, and work-life balance.
• Benefits vary by country and may include healthcare coverage, parental leave, retirement plans, relocation support, wellness programs, meal and transportation allowances, and other location-specific perks.