At Livi-Pro, I built the evaluation framework for a medical LLM assistant from scratch and helped ship it to production. I calibrated LLM judges against human-labelled conversations and added deterministic CI gates, regression scenarios, and price-grounding checks.
I also work as Lead Expert on Practicum's Data Analysis and Data Science tracks, where I review ML projects for data leakage and validation quality. I designed the 125-hour “Python for Data Analysis” curriculum and review criteria for the expert group.
Before that, at Contec, Consulting & Legal Services, I conducted forensic financial analysis for disputes and investigations. I reconstructed financial records, traced transactions, and produced analyses used to support claims in court.

