Senior AI Platform Engineer – LLMs & Agents

• Own the agent architecture and core AI platform
• Build a production-ready AI agent platform based on PydanticAI, LiteLLM, and Langfuse
• Develop RAG pipelines with pgvector at scale, including multi-tenant isolation, index rebuild strategies, hybrid search, and re-ranking
• Build a multi-tenant token quota service and tenant-specific evaluation frameworks
• Integrate MCP tools as a core part of the architecture
• Ensure end-to-end auditability from prompt to output
• Make architectural decisions for banking, insurance, and pharmaceutical clients
• Document relevant architectural decisions in ADRs
• Provide visibility into LLM calls, costs, and latency in Langfuse
• Report to Christoph Knöll (Founder/CEO) and work closely with the Head of AI
• Describe an LLM or RAG system you have brought into production as part of the application process

• 5+ years of production software engineering experience with Python
• Experience with LLMs in production, including tokens, latency, costs, and evaluations
• Hands-on experience building and operating RAG systems, including retrieval tuning and re-ranking
• Experience using an agent framework such as LangGraph, PydanticAI, or a comparable framework in production
• Excellent German and good English skills
• Experience sizing pgvector for large-scale data volumes
• Foundational knowledge of Kubernetes
• Experience with regulatory requirements such as the EU AI Act, DORA, BaFin/MaRisk, or GxP
• Experience with LiteLLM, Langfuse, or MCP is a plus but not required
• Based in Germany

• Permanent employment contract
• 40 hours per week
• 28 days of annual leave
• Fully remote, with residence in Germany
• Flexible working hours
• Team days at the Munich office approximately once a month
• Modern technology stack and dedicated time to evaluate new models, frameworks, and retrieval approaches
• Conferences and training
• Opportunities to represent the company externally as an expert, for example through talks or technical articles