Platform Engineer

About the role

TL;DR: This is not a regular engineering job. We're looking for exceptional builders who help us design and ship systems that define how the world's largest companies operate in the age of AI.

Who We Hire

We are relentless builders, and we're looking for people who share that obsession with shipping to customers.

As our founding Platform Engineer, your role is to build and scale a reliable, resilient, observable, performant and secure platform that makes our product engineers faster. We care about outcomes: you should be excited by building the foundations that let us ship a killer product fast, not by hard infrastructure problems for their own sake (though you'll have plenty of those).

You'll stand out if you

• Have owned infrastructure - networks, databases, deployments - through a period of real scale at a fast-growing startup or scale-up.
• Treat infrastructure as code by default (Terraform or similar) and stay on top of the latest platform tooling
• Simple, well-reasoned system design. Solutions that solve the problem and ship fast, rather than the perfect setup on a six-month roadmap
• Have informed, strong opinions on the big calls - like whether to stay on Postgres or move to something like Supabase or PlanetScale
• Understand how AI changes infrastructure - LLM gateways, caching, flaky providers, agent sandboxes, durable execution. You don't need to be an AI researcher, but you've played with the tools and have a point of view
• Build clean abstractions and paved paths that other engineers actually want to use

We look for exceptional talent, regardless where you spike. Experience level matters less than the quality of your thinking and the evidence of what you've shipped; building and scaling platforms with modern tooling, rather than just maintaining them.

We hire the best to join our team in New York & London, wherever they come from (we sponsor visas and provide relocation support!). We're convinced diverse teams ship better products and especially encourage underrepresented groups in tech to reach out.

What You'll Own

Everything that sits below the product code. Today that work is shared across the team; you'll own it, with the autonomy to set the direction, decide what matters most, as well as scope and prioritise work yourself with the backing of the broader engineering team. As our engineering team grows fast, every improvement you make compounds across everyone building Conduct.

• Deployment: fully automated, repeatable deployments of the whole platform, so we can stand up new regions - starting with the US.
• Enterprise-grade security: customer-provided encryption keys, network boundaries and tenant isolation that hold up to the scrutiny of the world's largest companies
• Bringing core services in-house: running durable execution, vector search and agent sandboxes inside our own infrastructure rather than renting them as a service
• Data: scaling and tuning our databases, and owning where our data layer goes next
• Developer experience: a fast, reliable local setup, preview environments from any branch, and build tooling and CI that keep up with a growing team
• Observability: end-to-end tracing, structured logging and alerting that product engineers can self-serve
• AI infrastructure: our LLM gateway, model routing, cost monitoring and resilience to flaky providers

Our Stack

Our stack prioritises what matters: shipping to customers-fast-without compromising on quality and engineering.

Fundamentals are more important to us than familiarity with any technology or tool: if you're strong, we're sure you'll pick it up fast.

Here's our setup

• Azure, managed with Terraform: everything as code, with multi-cloud integrations for LLM providers across AWS, GCP and others
• Kotlin and TypeScript services in a containerised monorepo, with CI/CD on GitHub Actions
• Temporal and PortKey: durable execution and enterprise-compliant LLM routing at the core of our platform
• Postgres for our data layer, with TurboPuffer for vector search
• Agent sandboxes running on Daytona and other providers
• OpenTelemetry and Honeycomb for observability