(Senior) AI Engineer (m/f/d) – Remote

Job description

What we do MAIA is an AI platform for companies. With MAIA, teams search their internal data and embed agents deep into their processes. We help companies use AI not just in individual departments, but across the entire organisation. On a technical level, this means: we build a GenAI application with our own agent systems, RAG pipelines and agent-based workflows, connect our customers' external systems and analyse their data. We use the best available models, operated in the EU, and make them reliable for demanding work with large, complex datasets. Many industrial companies have been using MAIA intensively for years. Following our successful 2026 funding round, we are now taking MAIA to a broader market. MAIA has been developed in Leipzig since 2021. What you'll do As an AI Engineer, you develop MAIA further and help decide from the start how we solve problems. What you can work on: From concept to feature. Together with product management, you sharpen rough concepts and own features from prototype to continuous improvement based on feedback and usage data. Chat and agents. You shape how people work with MAIA: asking questions, providing context and letting agents complete multi-step tasks, with traceable progress and verifiable results. Unlocking company knowledge. You connect external systems and make knowledge from many different file formats usable. This way, MAIA finds the right passages, delivers answers with sources and respects existing access rights. Models and quality. You integrate new language models and compare quality, cost and speed. With evals, you detect errors and systematically improve answers and agents. Secure, fast and stable. You build isolated environments for running code and tools and improve response times and data processing, so that MAIA runs reliably even with large knowledge bases. What we're looking for Solid software engineering. You have built and shipped backend systems and operated them in production over a longer period. Architecture, API design, data modelling and asynchronous processing are familiar to you. You recognise weak architecture even when a coding agent wrote it. Very strong in at least one of our languages. TypeScript, Python or Go, ideally one of the statically typed ones. You don't need all three. You'll get up to speed in the others. At home in the LLM ecosystem. You have built real applications with LLM…