How would you validate that an AI model is robust and reliable enough for national security use at Sandia National Laboratories?
