Data Scientist

Build and run an evaluation-driven LLM system for reliable responses and robust agent behavior using locally hosted language models. Define “good” for non-deterministic outputs, create regression and golden-set harnesses, and design LLM-as-judge and retrieval evaluation. Also develop production forecasting, anomaly detection, and early-warning signals from operational telemetry, partnering with data science, ML/inference, frontend, and product under enterprise security constraints.