Senior AI Engineer (m/f/d) – Agentic AI & Backend

What this is about

We're building MAIQ — the Legal Operating System for law firms. AI-native, designed from the ground up for German legal and data-protection requirements, and built to genuinely improve the day-to-day operations of legal professionals rather than add yet another tool on top.

Our goal is not to bolt individual AI features onto existing processes. We're building a system in which agentic AI, deterministic workflows and human-in-the-loop mechanisms work together to intelligently orchestrate a substantial part of a law firm's daily work.

We're a small, fast team from Cologne and Stuttgart, working together remotely and hybrid. Serious quality standards, a clear product vision, short lines of communication, a lot of heart. We already have secure infrastructure, an MVP in real use, a network of law firms for co-creation and first funding, so we can now hire our first employees. You're joining early enough to substantially shape the user-facing product layer — with a real share in the success through our VSOP.

Your role

You take on central responsibility for our AI backend and our agentic architecture:

• —You develop and own our Python-based AI backend end to end – from architecture decisions to production operation.
• —You build and evolve our agentic and multi-agent systems and make deliberate decisions about when an agent, a deterministic workflow or a combination of both is the right solution.
• —You work intensively with LangGraph and design robust state, routing and orchestration logic for complex processes.
• —You evolve our RAG and retrieval architecture and integrate vector databases and other knowledge sources into production workflows.
• —Together with the team, you make sure our LLM systems are observable, evaluable and debuggable – including tracing, evals, telemetry and clean LLMOps.
• —You bring AI pipelines reliably into production and help evolve our CI/CD and MLOps processes.
• —You think beyond your own backend: if a feature needs a GraphQL resolver or a Next.js/TypeScript route adjusted, that doesn't block you.
• —You translate product problems into sound technical solutions. You don't just implement requirements – you question, prioritise and propose the better way.

What you bring

• —Several years of experience in a senior role with a strong focus on AI and backend development.
• —Solid, production-grade Python skills – you have owned backend architecture yourself, not just built individual features within it.
• —Hands-on experience building agentic or multi-agent systems that actually run in production – not just as an experiment or proof of concept.
• —Direct hands-on experience with LangGraph and working knowledge of LangChain and/or LlamaIndex. We care less about framework trivia than about your judgement of which pattern makes sense when.
• —Experience integrating and operating a vector database in a live system.
• —Confident use of LLMOps tools such as LiteLLM, Langfuse or comparable solutions for tracing, evaluating and debugging production LLM applications.
• —Solid MLOps understanding: you have already put a model or an LLM pipeline into production and run a working CI/CD pipeline for it.
• —You're technically broad enough to read and extend a GraphQL resolver or a Next.js/TypeScript route when needed – without frontend development having to be your focus.
• —You want to work closely and directly in a small team: daily exchange with the founders and the engineering team, fast decisions instead of corporate processes, and a lot of responsibility even when not everything is defined yet.
• —You're up for a real transition period side by side with your predecessor, to take over knowledge, architecture decisions and context properly.
• —German and English at conversational level (around C1, no certificate needed) – you can hold fluent conversations in German. Our documentation and code are in English.

What makes you stand out

Nice to have, not a requirement:

• —Experience in cloud engineering, ideally with Microsoft Azure. Our infrastructure runs on Azure, including Container Apps, Cosmos DB, Blob Storage and Service Bus.
• —Broader full-stack experience with Next.js and TypeScript that goes beyond reading code and making smaller changes.
• —Startup experience – you know small teams, fast iterations and priorities that can shift.
• —Experience with evals, guardrails or human-in-the-loop architectures for production LLM systems.
• —An appetite for a regulated domain in which quality, traceability and data protection are part of the product.

What we offer

• —Ownership from day one: you take over a central technical area of our product and can shape architecture decisions directly.
• —AI at the core of the product: you won't be bolting a chatbot feature onto an existing SaaS product. Agentic AI and intelligent orchestration are part of MAIQ's fundamental architecture.
• —Technical freedom: we don't want to use as many frameworks as possible, but the right ones. Good technical arguments count for more than existing conventions.
• —A real handover instead of being thrown in at the deep end: during the transition you work directly with your predecessor and can take over architecture, codebase and decisions in a structured way.
• —Speed + quality: we ship fast, but we're building a product for a demanding, regulated audience. Both, not just one.
• —Fair startup salary plus an attractive VSOP: you share in the success you help build.
• —Permanent position, hybrid work, living in or around Cologne or Stuttgart. 1–2 working days per week in the office and regular in-person team meetings matter to us.
• —A modern stack built around Python, LangGraph, LLMs, Next.js/TypeScript and Azure.
• —Short lines of communication, direct collaboration with the founders, real influence and an experienced team with high technical and product standards and a passion for innovation.