Dr. Mark Hays provides professional consultation with industry partners during the summer on how to write and maintain AI-enabled code safely in risk-adverse environments. Dr. Hays has experience in the intersection of traditional software quality assurance, NLP, and statistics-based AI systems, and a strong foundation for thinking about LLM-based AI systems from a quality perspective. Since the rise of LLMs, Dr. Hays can work with you to explore LLMs capabilities and especially weaknesses to insulate software quality and development efforts from their pitfalls.
1. Helped a Wisconsin-based company in 2025 research a new change management process to safely apply LLMs to update a legacy invoice processing system from an obsolete programming language to a contemporary programming language. The research was scoped to the oldest core of their legacy system. Delivered my revisions to the codebase to facilitate their efforts and gave a company talk on the nuances of my newly researched process.
2. Helped a Wisconsin-based company in 2020 evaluate whether their quality assurances processes of AI-enabled systems were adequate. We scoped the effort to a system designed to an AI system that helped nursing homes plan capital expenditures related to maintenance. I delivered both my analysis of their system and a custom tool they could use to generate their own quality reports.
3. Helped Indiana-based Indigo bioAutomation in 2015 identify parts of their core AI-enabled "Ascent" product that were most at risk to bugs. Engineered a custom code quality inspection tool to run on Ascent that could highlight problem areas where testing needed to be bolstered. Delivered the custom tool by giving a talk to their Quality Assurance team on how to use the tool and linked the code for the tool itself.