Laura H. Maza
Software engineer working in risk and AI governance for highly regulated industries · Independent AI Researcher
I work where AI systems meet accountability. For more than twenty years I have moved between language, software and compliance: as a professor of English language and literature, a software engineer and CTO, and a governance and compliance lead in regulated industries.
Today that becomes two lines of work that feed each other. In practice, I help organizations govern risk and adopt AI under frameworks like ISO/IEC 42001 and the NIST AI RMF. In research, I measure how large language models react when the wording of a prompt changes but its meaning does not.
Professional practice
Governance, risk and compliance; AI governance; regulatory QA; technology leadership from engineering to the executive table, through MAZALab.
See areas of practice and credentialsResearch
Controlled experiments on LLM sensitivity to voice, clause structure and register, plus open-source tools for evaluating model quality.
See research and projectsIn a first, small-scale experiment, shifting a prompt from formal to informal register moved a model’s next-token distribution about fifteen times more than switching active to passive voice. Read the preliminary result