The impact you will create:

• Architect, implement, and iterate on complex prompt logic, agentic decision flows, and multi-agent routing graphs.
• Design tool schemas, API integrations, and structured outputs for LLM reasoning and execution.
• Implement OpenTelemetry tracing and observability across agent pipelines, LLM calls, and tool interactions to track decision paths and performance.
• Build robust context management and retrieval-augmented generation (RAG) strategies to ensure accuracy and low latency.
• Handle AI error modes and edge cases, designing fallbacks and safety guardrails for when models hallucinate or are uncertain.
• Conduct continuous evaluation and testing of LLM behavior, aligning outputs directly to business logic and user intents.
• Embed traceability and logging into AI decision paths to support auditability and compliance (e.g., EU AI Act).
• Collaborate with Data Scientists (metrics/evals) and DevOps/Data Engineers (infrastructure) to deploy scalable agentic systems.

Experience and skills you will need:

• Demonstrated software engineering background combined with deep hands-on expertise in LLMs, prompt engineering, and agent frameworks (e.g., LangChain, LlamaIndex, AutoGen).
• Non-negotiable: Strong business and domain understanding with the ability to translate ambiguous business goals into behavioral LLM design.
• Practical experience with model routing, cost/latency tradeoffs, and failure-mode mitigation in production GenAI applications.
• Proficiency in Python, API development, and software development practices (Git, unit testing, CI/CD).
• Demonstrated ability in stakeholder communication and management.
• Good understanding of related data disciplines (e.g., Data Engineering, Data Science, MLOps).
• Ability to understand, analyze and translate business requirements into data-driven solutions.
• Enthusiasm for collaboration, knowledge sharing, and joint learning in a fast-evolving AI landscape.
• Very good English skills — German is a plus