At IBM, I develop Python evaluation infrastructure for assessing LLM-generated code against programming tasks and benchmark datasets. I build reusable test harnesses and benchmark runners that validate outputs, classify failures, and produce reproducible results.
Previously, I developed Python automation components at Toyota Boshoku America for model and benchmark evaluation, then supported backend, API, and React application work at American Express. My work has also included SQL-backed workflows, automated testing, and CI/CD integration.

