
Principal Data Scientist - Agent Builder
Lead technical direction for evaluating, improving, and scaling conversational/agentic search quality (RAG, agents, retrieval, citations, vector search). Define evaluation strategy (offline/online, metrics, LLM-as-judge), prototype and productionize evaluation pipelines and telemetry, influence product and model decisions, and mentor other data scientists and engineers. Requires 8+ years experience and hands-on skills with Python, PyTorch/Transformers, retrieval methods, and Elasticsearch.