Junior Product Engineer

Architect Turned Software Engineer

This role is for people with real-world construction experience, ideally as an architect, who are now building software. You don't need years as an engineer, but you do need real evidence of it, courses, projects, or work where you've actually built and shipped something. The construction experience is required; we grow the engineering with you.

The job of an architect is hard. Accountable for multi-million-euro projects, they wear every hat, from creative draftsman to material specialist to Excel professional, and lose evenings and weekends to backoffice work legacy software never solved.

Our mission is to give architects an easy-to-use solution so they spend their time building homes instead of fighting Excel. We're building our first true AI products to do exactly that: turning the mess of a real planning office (site photos, Whats App voice notes, scattered Excel, CAD, email) into the formal deliverables architects are legally on the hook for, like LVs per DIN 276, HOAI fee proposals, construction diaries, and defect records.

You've lived that pain. You know what these deliverables actually have to get right, because you've been the one on the hook for them. Now you'll build the software that fixes it.

Most "AI-native" jobs bet on a product that doesn't exist yet. This isn't one. We're serious software for German planning offices, with real users and real revenue, now building our first true AI products on top. The AI features themselves are still early here, which is the rare part: a greenfield mandate inside a company that already works, on a codebase that ships to paying customers every week.

This role exists because the best person to build for architects might be an architect. AI tooling now makes it realistic for someone with deep construction expertise to become the engineer who ships the fix, and that's exactly the bet we're making with you. We keep the team small on purpose (a few product engineers, a designer deep in product, an incoming product lead, the founders), so what you build is felt by everyone immediately.

Backed by some of Europe's best B2B SaaS investors. Remote-first, with a Berlin office.

You come in with construction expertise we can't teach; we grow the engineer on top of it. The trade runs both ways: we teach you to build, you make the whole team faster and sharper about the domain.

• Build real AI features, with support. You'll own meaningful pieces from the start, ship to real users, and grow toward full end-to-end ownership as your engineering deepens.
• Be the domain in the room. You already know what an LV, an HOAI proposal, or a defect record has to get right. You translate that into what we build, and you catch where the AI is subtly wrong in ways an outsider never would.
• Tighten the loop between domain and build. We get domain insight from user calls and visiting architects on-site, like every product engineer here. What you add is a much faster loop: you understand your peers' problems quickly because you've had them, so the distance between "what the domain needs" and "what we ship" gets shorter.
• Learn to engineer for real. Fundamentals, system thinking, and the judgment to assess your own output, using AI to accelerate your understanding, not to skip it. We invest in this deliberately.
• Get rigorous about whether the AI is good. Help build and read evals, grounded in domain truth you can actually judge.

You'll be mentored by senior product engineers, with a real path to grow into a full Product Engineer role over time. This is a full-stack product role across our Type Script (Node backends, React frontend) and Java/Spring Boot stack on AWS and Postgres.

• Construction is your edge. Architect, Bauingenieur, or contractor, this is what you bring that we can't teach a normal junior.
• You've already taught yourself to build, and it shows. A bootcamp capstone, automations for your old firm, real side projects, even heavily AI-assisted, but something that works and that you can walk us through. We bar on evidence of building and aptitude, not on a CS degree or years of experience. We can't take an absolute beginner here.
• AI already helps you learn. You reach for coding agents and LLMs to understand established patterns, practices, and complex problems faster. The point isn't how productive it makes you; it's that it accelerates your understanding instead of replacing it, so you can always reason about and assess what you build.
• High agency and a fast, curious learner. You take a vague problem, go get the missing context yourself, and turn it into something shipped. You don't wait for a clean ticket.
• Genuinely want to become an engineer. Not a product advisor who can code a little. You want system thinking and real fundamentals, and you'll put in the work to get them.

Required:

German at B2 or above (our working language is English, but our domain and ground truth are German). Strong plus: the deeper your…