<div class="show-more-less-html__markup show-more-less-html__markup--clamp-after-5 relative overflow-hidden"> <strong>Job Description<br/><br/></strong><strong>Founding Full-Stack Engineer | AI Infrastructure Startup<br/><br/></strong><strong>Munich (Hybrid)<br/><br/></strong><strong>€70,000 – €90,000 + Equity<br/><br/></strong>A VC-backed early-stage startup is building an AI-native operating system for one of Europe’s largest and most inefficient industries.<br/><br/>Today, billions are lost because critical project knowledge lives across PDFs, meeting notes, emails, and fragmented systems. Every major project starts from zero.<br/><br/>They’re solving that by building an AI co-worker that sits at the centre of complex operational workflows - turning scattered information into structured execution, proactive decisions, and compounding institutional memory.<br/><br/>The product is already live inside major large-scale infrastructure programmes with real operational consequences when things go wrong. Every deployment improves the system, creating a long-term moat competitors can’t easily replicate.<br/><br/>Backed by top-tier early-stage investors, including specialist funds with deep exposure to infrastructure, AI, and some of Europe’s strongest founder ecosystems.<br/><br/>The Role<br/><br/>This is a genuine founding engineer opportunity.<br/><br/>You’ll work directly with the founders and own major parts of the product from day one - helping define architecture, product decisions, and how AI should actually show up inside critical workflows.<br/><br/><strong>Two Major Challenges<br/><br/></strong><ul><li> AI-native workflow software, not chatbot theatre<br/><br/></li></ul>The goal isn’t another chat interface on top of documents.<br/><br/>It’s building software where AI operates through automation, background processes, alerts, workflows, and delegated execution - knowing when to act silently, when to recommend, and when humans should stay in control.<br/><br/><ul><li> Extending agent systems all the way to the frontend<br/><br/></li></ul>Agent infrastructure doesn’t stop at backend orchestration.<br/><br/>You’ll build frontend systems that handle streamed outputs, tool execution traces, partial renders, interruptions, and recovery flows as first-class product experiences - not UI afterthoughts.<br/><br/>Stack<br/><br/>TypeScript, Next.js, Vercel, Supabase (Postgres + vector search), modern LLM tooling, agent frameworks, observability tooling, and production-grade AI workflows.<br/><br/>Greenfield environment. No legacy systems.<br/><br/>Strong tooling culture with dedicated AI experimentation budget per engineer.<br/><br/>What They’re Looking For<br/><br/><ul><li>Strong full-stack engineering ability with excellent product instincts</li><li>You think from user pain ? product decision ? measurable outcome</li><li>Comfortable speaking with customers and understanding operational workflows</li><li>Fast execution without sacrificing reliability</li><li>Strong ownership mindset and comfort operating in ambiguity</li><li>You can prototype quickly, test assumptions, and improve based on real usage</li><li>Bonus points for startup experience, strong open-source work, AI systems, document understanding, or systems that replaced manual operational work<br/><br/></li></ul>Level is flexible - they care more about exceptional people than rigid seniority labels.<br/><br/>English-speaking environment. German is helpful but not required.<br/><br/>Why Join<br/><br/><ul><li>True founding engineer ownership</li><li>Direct access to founders and strategic decisions</li><li>Immediate production impact on high-value real-world problems</li><li>Strong equity package (0.5%–1.5%)</li><li>Fast learning curve with frontier AI systems</li><li>Hybrid setup: 3 days/week in central Munich office</li><li>Clear path into senior IC or future leadership roles depending on your strengths<br/><br/></li></ul>Interview Process<br/><br/>Fast and simple. No endless loops.<br/><br/><ul><li>Technical first stage</li><li>Founder conversation</li><li>Final onsite case study in Munich</li><li>Decision within 24–48 hours</li></ul> </div>