Key responsibilities: - Build or configure production-ready agents with memory and state for clinical and operational workflows; own them end-to-end including orchestration and integration. - Support deployment and operational troubleshooting; define and maintain an offline evaluation harness; combine LLM-as-judge with deterministic checks and integrate regression gates into CI. - Design and optimize retrieval augmented generation (RAG) pipelines over structured and unstructured clinical text; measure retrieval quality (faithfulness, relevance, recall). - Define acceptance criteria, failure-mode analyses; design guardrails, human-in-the-loop processes, and escalation paths; provide validation and traceability under Medical Device Regulation MDR and the EU AI Act. - Instrument production observability: trace LLM and tool calls, retrieval steps; monitor quality drift; turn production failures into new evaluation cases. - Optimize cost and latency: apply model routing/selection, caching, token budgeti

Key responsibilities: - Build or configure production-ready agents with memory and state for clinical and operational workflows; own them end-to-end including orchestration and integration. - Support deployment and operational troubleshooting; define and maintain an offline evaluation harness; combine LLM-as-judge with deterministic checks and integrate regression gates into CI. - Design and optimize retrieval augmented generation (RAG) pipelines over structured and unstructured clinical text; measure retrieval quality (faithfulness, relevance, recall). - Define acceptance criteria, failure-mode analyses; design guardrails, human-in-the-loop processes, and escalation paths; provide validation and traceability under Medical Device Regulation MDR and the EU AI Act. - Instrument production observability: trace LLM and tool calls, retrieval steps; monitor quality drift; turn production failures into new evaluation cases. - Optimize cost and latency: apply model routing/selection, caching, token budgeti