Independent AI safety researcher. Built BiasClear, an open-source tool for detecting structural persuasion in LLM outputs. 20+ years in banking operations and risk management. Published: Persistence of Information Theory (DOI 10.5281/​zenodo.18676405).