In the modern battlespace, algorithmic outputs are increasingly driving tactical and strategic decisions. But how do we know the machine is actually working? Drawing on decades of analytics applied to military problems, Walt DeGrange and Dave Saranchak provide a unique perspective on how AI can be used effectively in today’s ever-changing environment. This three-day intensive course cuts through the theoretical noise of Artificial Intelligence (AI) and provides what AI is and is not and how to make sure it is doing what you want it to do. The course also directly addresses the 'Validation Challenge' — the critical friction between unclassified model development and the high-stakes, classified operational environments where commanders must make decisions.

Designed specifically for Military Analytics Professionals, this course explores the foundational mechanics of Machine Learning, Deep Neural Networks, and Foundation Models. Rather than just teaching students how to run a Python script, this curriculum demands they challenge the model. Guided by Dr. Richard Feynman’s philosophy of rigorous questioning and practical experimentation, students will learn to break algorithms down, test their operational boundaries, and expose their vulnerabilities. Throughout the course, professionals will work on an overarching project utilizing Python and local Large Language Models, culminating in a critical exercise: translating complex AI constraints into actionable, straightforward briefings for military leadership.

Overarching Course Project: Operation Red Pill:

Throughout the three days, students will be grouped into analytic cells. They will be tasked with deploying a local Large Language Model and developing a predictive classification model using Python. Emulating Feynman's experimental mindset, they will intentionally introduce adversarial noise into their data, formulate hypotheses about how the models will fail, and run controlled experiments to test those hypotheses. On Day 3, each cell will present their findings, explicitly explaining the pros, cons, and operational constraints of their models to a simulated panel of senior military leaders.