Forward Deployed Engineer (m/f/d)

Job description You own the technical outcome of what you build. You sit next to the customer problem with an AI Product Builder, shape the requirement together, then implement it — yourself and with coding agents. You are not a spec-taker. If the spec will not serve the outcome, you push back. You choose tools, frameworks, and architecture for your scope. You treat AI as the main way work gets done, and you still own craft: code quality is yours whether you typed it or an agent did. Your tasks Own a product or a major feature area end to end: discovery, implementation, quality, and delivery Turn intent into specifications precise enough for coding agents to execute, then orchestrate and review their output Decide where AI judgment is acceptable and where the system must be deterministic Judge work on outcomes and tests, not on whether the implementation looks clever Tell “it passes” from “it is production-ready” Co-design the solution with the AI Product Builder as a near-peer. Challenge ambiguous or weak specs instead of building them literally Identify and assess AI use cases in your area on value, feasibility, and risk — and stop the ones that do not hold up Design and run evaluation for AI-delivered work: test cases, inconsistency, hallucination, and bias Raise responsible-AI, privacy, and regulatory concerns early. Follow the governance that applies to what you ship Work with Engineering on integration constraints so what you ship fits the systems it has to live in Your profile Critical thinking, proactive problem solving and innovation are key Strong interest and hands-on experience with AI tools (e.g. Claude, Cursor, MCP), different LLM models and prompting Proficiency in TypeScript or other languages such as Python, Java, or C# Proven ownership of a feature area or product slice: you have decided implementation and architecture from the problem, not from a ticket Fluency directing AI coding agents: you write specs and acceptance criteria, supervise execution, and reject output that is not production-ready Enough product and customer judgment to challenge a spec that is clear but wrong A personal standard for code quality and craft, including on AI-generated code Comfort working with product, engineering, and governance counterparts without waiting to be handed a perfect brief Fluent German (C1 level) and English (B2 level) proficiency are required Educational background in Computer Science or a relational discipline is preferred Nice to have Experience with RAG pipelines, agent frameworks, vector stores, or orchestration — we grow this with seniority A track record of cutting rework by making specs unambiguous before agents run Familiarity with responsible-AI practice and regulation relevant to production software in Europe (including the EU AI Act) You have evaluated third-party AI components or platforms, not only used them Your benefits Work across AI-powered SaaS and cloud-native products, use modern AI…