Proven experience building production systems around LLMs, especially fine-tuning and prompt engineering Strong backend engineering skills in Python (FastAPI or similar) Ability to build service layers and interface with external APIs (e.g. OpenAI) Practical experience with vector databases like Pinecone, Weaviate, Qdrant, or FAISS Experience tracking and analyzing user behavior / feedback to extract ranking signals and patterns Familiarity with embedding models, chunking strategies, and hybrid search logic Ability to design and implement lightweight data pipelines for dynamic preference updates and context generation