数据策略研发工程师（医疗大模型 Agent 方向）（J106671）
Core
Design and implement memory strategies and foundational content strategies for medical industry Large Language Model (LLM) Agents to enhance multi-turn dialogue, long-term task continuity, and precise medical knowledge retrieval.
Role type
Senior IC data strategy engineer (medical LLM Agent)
Builds
Multi-layer memory systems (long-term, short-term, persona), structured medical content databases, and evaluation frameworks for Agent performance.
Domain
Healthcare + Large Language Models (LLM) + Agent Systems
Deliverable
production ML models
Required skills
Memory management, Context engineering, RAG, NLP (entity extraction, semantic matching), Python, Deep learning frameworks (PyTorch), Statistical analysis, A/B testing, LLM-as-Judge evaluation
Preferred skills
Medical knowledge graph construction, Medical domain expertise, Planning/Reflection/CoT reasoning paradigms, Vector search, Prompt engineering, Model fine-tuning
Technologies
Python, PyTorch, Vector retrieval, LLM frameworks
Responsibilities
Design memory strategies for user long-term/short-term memory and scenario personas; Build structured content systems for doctors, hospitals, diseases, drugs, and treatment plans; Process user queries and interaction data to support strategy optimization; Establish full-link data and evaluation systems to quantify strategy impact; Collaborate with product and algorithm teams to scale strategy deployment; Track and innovate on LLM/Agent technologies for health scenarios.
