Full-Stack AI Engineer
Mission
Construction is a $13T industry that still runs on PDFs, email chains, and manual cross-referencing. Every major project generates thousands of pages of contracts and commercial documents that nobody has the time to read properly — and mistakes cost millions. We're building the intelligence layer that changes that.
What You'll Own
Data & Infrastructure: Own our FastAPI / Supabase stack — schemas, ingestion pipelines, and the processing layer that turns messy construction documents into structured data.
AI Agents: Build the multi-agent system powering bid comparison, clause extraction, risk detection, and knowledge graph sync. Decide how agents collaborate, fail, and improve.
Evaluation: Define evals that actually measure whether the system works — accuracy, recall, failure cases, cost.
Product Interface: Build the Next.js front-end contract managers trust enough to base award decisions on — not a chatbot, a structured workspace.
Production: Deploy, monitor, keep it running. Observability, CI/CD, and the production discipline that earns enterprise trust.
Who We're Looking For
Not a "fullstack developer who's interested in AI." Not someone who needs a ticket to know what to do next.
We're looking for someone who cares deeply about their craft — the kind of engineer who builds eval infrastructure before anyone asks for it, who notices when an abstraction is off and fixes it, who takes pride in clean systems.
You'll stand shoulder-to-shoulder with the CTO as co-builder. You set the engineering bar — and you keep it there.
Munich-based, or ready to relocate.
Technical Skillset
You've built and shipped AI-powered products with LLMs in production — not just prototypes, not just demos
Strong Python, hands-on FastAPI or equivalent
RAG pipelines: embeddings, vector stores, retrieval strategies — you've done the hard parts
Experience with LangGraph, LangChain, or similar orchestration frameworks
You've defined and run real evals for AI systems — you know vibes-based testing breaks in prod
Software engineering fundamentals that don't need explaining: clean APIs, tests, versioning, security
You think in systems — balance speed with scale, design things that evolve instead of collapse
Commitment & Mindset
High agency. You see a problem, you fix it. You don't flag it in Slack and wait.
Product-obsessed. You think in user outcomes. You've closed your laptop and thought "would I actually use this?"
Energized by intensity. We're building fast, in-person, every day. We want people who thrive in that environment.
Benefits
Impact that's real. Construction is one of the largest, least-digitized industries on the planet. We're not optimizing ad clicks.
Whatever tools you need. Hardware, Claude Code, unlimited AI credits, productivity setup — just ask.
Skin in the game. Competitive comp + equity. You shape the product, the architecture, the culture.
Intensity + recovery. We push hard and enjoy it. Team dinners, EGYM Wellpass, the right balance.
Occasional San Francisco offsite.
About the interview process
We're thorough but we respect your time.
Intro call — Meet a founder. (~30 min)
Technical deep dive — We go deep on your experience as a developer and as a person, how you think about systems, and how you'd approach real problems in our domain. (~1 hr)
Take-home — A time-bounded, practical challenge based on our actual work. AI + full-stack, end-to-end.
Technical Follow-up — Present your take-home, walk us through your decisions, and dig into the trade-offs together.
On-site day — Work with the full team from our Munich office at Sendlinger Tor. See how we actually build.
Offer.