If physics could be unified through geometry, why not learning? I'm a researcher based in Toronto investigating how singular geometry shapes what neural networks can learn, what representations they develop, and how belief geometry constrains action and planing. Symmetries in data and models create singularities—not pathologies to avoid but essential structure that governs learning, generalization, and decision-making. I'm developing mathematical foundations using singular learning theory, algebraic geometry, and stochastic analysis—working toward a unified understanding of how intelligent systems learn, represent, and act. Deep learning is ready to transition from art to science. The foundations are within reach.