Full Job Description
About the role
Brainbase is a platform of platforms: we support every publicly available harness, frontier and open-source models, and the primitives teams need around them (routing, cost controls, evals, guardrails, auth, permissions, sandboxes, payments). As a Senior MTS, you'll own large pieces of that platform end to end, from design to production to the customer who depends on it.
You'll work across our core primitives and the infrastructure that runs them. Our services run on our managed cloud, on AWS, Azure and GCP, and in sovereign on-premise deployments, with low latency and high reliability. Some are simple containers (our auth service); some are hard to control and scale (sandboxes). You'll decide how a lot of this gets built.
What you'll do
• Own one or more core primitives (for example routing, evals or sandboxes) and be the person who decides where it goes next
• Design and ship infrastructure that runs the same way across our cloud, the major clouds and on-prem
• Work directly with customers and our forward deployed team when a deployment pushes the platform somewhere new
• Set the engineering bar: code review, architecture, on-call, and how we ship
Who we're looking for
• Several years building and running production systems, ideally distributed systems or developer infrastructure
• Excellent with TypeScript and Python; comfortable with containers, Kubernetes and at least one major cloud
• Real experience building with LLMs and agents, beyond calling an API
• A record of owning things end to end: a system you designed, shipped and kept running
• Experience managing a team, ideally at a startup
• You do not write code manually
Nice to have: experience with sandboxing or isolation (Firecracker, gVisor and similar), on-prem or air-gapped deployments, model serving, or a past startup.
Don't check every box? Apply anyway. We're open to meeting anyone exceptional, even if you don't meet all of the requirements above.
Your first month
• Week 1: Ship a fix to production, own an adjacent project from 0 to 1, like building a new platform for our partners to use. You'll also use Brainbase to make a new agent for the team, like a Slack bot for making our Doordash orders.
• Weeks 2-3: Take ownership of one primitive, feature, optimization, or piece of infrastructure from our sprint, and ship it.
• Week 4: Spec and own your own improvement to our platform.