I lead independent applied-AI research through the DUJ-FEITH Foundation, a Canadian nonprofit focused on governed AI execution, agent safety, multi-agent stability, and edge validation.

My work examines a practical question: when an AI system proposes an action, what determines whether that action should be permitted to become real?

I design and test policy gates, evaluation harnesses, simulations, and vision-to-action pipelines, including physical-sensor validation on NVIDIA Jetson Orin Nano. Selected methods, experiments, and results are documented publicly on GitHub.

My background spans more than 15 years in strategic partnerships, sales leadership, and market development. I connect technical experimentation with practical evaluation frameworks, ecosystem strategy, and responsible real-world deployment.

GitHub: https://github.com/IsabelleThivierge