Senior AI Platform Engineer – LLMs & Agents

This is a fully remote position, open to applicants in Germany.

📋 Description

• Take ownership of the agent architecture and the foundational AI platform.

• Develop a production-ready AI agent platform utilizing PydanticAI, LiteLLM, and Langfuse.

• Create RAG pipelines with pgvector at scale, ensuring features such as multi-tenant isolation, index rebuilding strategies, hybrid search, and re-ranking.

• Establish a multi-tenant token quota service along with tenant-specific evaluation frameworks.

• Integrate MCP tools as an essential component of the architecture.

• Guarantee complete auditability from prompt to output.

• Make architectural decisions tailored for clients in banking, insurance, and pharmaceuticals.

• Document significant architectural decisions through Architecture Decision Records (ADRs).

• Provide insights into LLM calls, associated costs, and latency metrics in Langfuse.

• Report directly to Christoph Knöll (Founder/CEO) while collaborating closely with the Head of AI.

• Share details of an LLM or RAG system you have deployed in production as part of your application.

⛳️ Requirements

• Over 5 years of experience in production software engineering using Python.

• Proven experience with LLMs in a production environment, including aspects like tokens, latency, costs, and evaluations.

• Practical experience in building and managing RAG systems, focusing on retrieval tuning and re-ranking.

• Familiarity with agent frameworks such as LangGraph, PydanticAI, or similar frameworks in a production setting.

• Proficient in German and possess good English communication skills.

• Experience in sizing pgvector for substantial data volumes.

• Foundational knowledge of Kubernetes is essential.

• Understanding of regulatory requirements including the EU AI Act, DORA, BaFin/MaRisk, or GxP.

• Experience with LiteLLM, Langfuse, or MCP is advantageous but not mandatory.

• Must be based in Germany.

🏝️ Benefits

• Permanent employment contract.

• 40-hour work week.

• 28 days of annual leave.

• Fully remote position for residents in Germany.

• Flexible working hours.

• Monthly team days at the Munich office.

• Access to a modern technology stack and dedicated time to explore new models, frameworks, and retrieval methods.

• Opportunities for conferences and training.

• Chance to represent the company externally as an expert, including speaking engagements or writing technical articles.