We are hiring Open positions

We’re also hiring humans

Square8 builds AI workers: agentic systems that hold down real jobs inside companies, follow the rules they’re given, and log every action they take. Our first worker is already live in production, and the people who build the next ones are the ones we’re looking for now.

Open roles

Both roles sit in the early engineering team and work directly with our CTO

AI Product Engineer
You’re a forward deployed engineer. Time to take it to the next level.
Human worker
Location
Zurich, Lisbon or London
Compensation
Cash plus share plan
The role
We’re building AI workers: agentic systems that work as teammates. You’d drive parts of the product end to end, from working out what a customer actually needs through to what ships and how it holds up in production. No product manager. You own what you ship.
Read the role
AI Platform Engineer
Build the harness that builds the harness.
Human worker
Location
Zurich, Lisbon or London
Compensation
Cash plus share plan
The role
Own how Square8 builds software: specifications, orchestration, verifier agents, and evaluation you can trust. Then own verification inside the product.
Read the role

AI Product Engineer

You’re a forward deployed engineer. Time to take it to the next level.

The job

Square8 builds AI workers: agentic systems that work as teammates. They have their own email address and Slack handle, follow your company’s rules, remember what a colleague would remember, and log every action they take. Our first worker is already live in production.

You’d join the early engineering team and own parts of the product end to end. You take a piece of what the customer actually gets: defined, built, shipped, and yours once it’s live.

If you’ve worked as a forward deployed engineer, most of this will be familiar. You still do the discovery. You still sit with a finance team until you understand how they actually work. The difference is what happens next. Instead of building a configuration for that one client, you build the product, and what you learned there goes into what every other customer gets.

There is no product manager here. What gets built is your call.

You own what you ship

We organise our work around pieces of the product: one integration, one part of the agent, one customer’s workflow. Each piece has an engineer driving it.

When you take one, you take all of it: the customer conversations that define it, the specification, the fleet of coding agents that builds it, the release, and how it behaves at 3am three weeks later. You pull in other members of the team for as long as the work needs them, and the group changes as the work does.

We all know which parts of the product are whose.

Is this you?

You’re a forward deployed engineer who has taken agentic systems into production, with customers in the room. You’re ready for the next level: turning that work into a product, so what you learn at one customer compounds across every other one instead of being rebuilt for the next. Less consulting, more product.

What you’ll do

  • Spend real time with the people whose work the AI worker will take over, and figure out what the job actually involves
  • Write specifications precise enough for AI agents to build from
  • Run fleets of coding agents across parallel workstreams, and decide how the system fits together
  • Build the core of the agentic system: coordinator, memory, rule engine, audit trail, and the specialist agents that drive SAP, Excel and documents
  • Get the first worker live at a customer within eight weeks

What we’re looking for

  • You’ve taken agentic systems into production, not just to a demo, with customers involved along the way
  • Experience integrating with enterprise systems: ERP, mail, documents, and the ones with no usable API and no documentation
  • Strong product judgement. You decide what gets built, and you can say why
  • You work AI-first. You’d rather run ten workstreams than write one of them yourself
  • At some point you’ve shipped something narrower than a customer asked for, because the narrow version worked for everyone else too

Nice to have: forward-deployed or applied AI engineering, or time as a founder.

What we offer

Compensation is a combination of cash and a share plan, worked out between you and the founders. You’ll work directly with our CTO, and with a real enterprise customer from your first week. Nobody will hand you a spec. This is an in-person role, based in Zurich, Lisbon or London.

When you apply: Tell us about something you built end to end: what generalised, and what didn’t.

AI Platform Engineer

Build the harness that builds the harness.

The job

Square8 builds AI workers: agentic systems that hold down real jobs inside companies, follow the rules they’re given, and log every action they take. Our first worker is already live in production.

We build them with a small team directing large fleets of AI coding agents. Whether that works comes down to the machinery around the agents: the specifications they build from, the verifiers that check their work, the tests that catch what the verifiers miss, and the evaluation that tells you whether things are getting better or just changing.

That machinery is your job. You’d own how Square8 builds software.

It also sits closer to the product than internal tooling usually does. Verification, audit and bounded autonomy are what we sell. Solving them for our own engineering is a first pass at solving them for customers.

Is this you?

You’ve built the harness other engineers reached for by choice. Now you want the bigger version of that problem: the machinery that lets a small team direct large fleets of AI agents and trust what comes back. That is where software engineering is going, and almost nobody has built it properly yet. You’d build it here, on a product where verification is the thing we sell.

What you’ll do

  • Own the specification format. Everything the agents produce depends on it
  • Build the orchestration: dispatch, isolation, parallelism, cost control
  • Build the verifier agents that check AI-written code, and measure honestly what they catch and what they miss
  • Build testing for systems that aren’t deterministic: replay, seeded runs, and flakiness you can act on
  • Make correctness something the system establishes, rather than something a person has to go looking for
  • Own verification inside the product itself, once you’ve solved it for us

What we’re looking for

  • You’ve built infrastructure that other engineers used by choice
  • Hands-on experience running coding agents at scale, and clear views on where they break
  • Comfort with statistics: catch rates, base rates, and knowing when a number isn’t good enough to act on yet
  • You’ll argue with our CTO, and we’d be worried if you didn’t

Nice to have: agent orchestration infrastructure, open evaluation frameworks, build systems at scale, developer-productivity work where you owned the metric.

What we offer

Compensation is a combination of cash and a share plan, worked out between you and the founders. You’d work alongside the CTO as a peer, with no layers in between. Direct work with a real enterprise customer from your first week. This is an in-person role, based in Zurich, Lisbon or London.

When you apply: Tell us about infrastructure you built that people actually used.

How we hire

We are a fast-paced company with a strong sense of urgency. You can expect the same from us in the hiring process.

01

A short first conversation

We check whether what you want to work on and what we need line up.

02

How you think about the system

A longer conversation about what we’re building: what you would question, and what you would expect to be hard.

03

A deep dive with the early team

You meet the whole early team and go deep on the work together.

Square8 is an equal-opportunity employer. If your background is unusual for this role, that may be exactly the point. Tell us why.

Build AI workers with us

Send us a short email: who you are, how to reach you, why you’re applying, and any links you want us to look at. No CV, no cover letter, no PDFs.

hello@square8.ai