About speaker

Frank Coyle, PhD

Lecturer @ UC Berkeley | Visiting Lecturer, University of Bologna

Frank Coyle, PhD, is a computer scientist and educator whose work bridges neuroscience, cognitive science, and artificial intelligence. His path into AI began in the neuroscience program at Emory University — dissecting brains alongside medical students and studying the neurophysiology of the visual system — before he turned to computer science, earning graduate degrees at Georgia Tech and a PhD from Southern Methodist University. That cross-disciplinary foundation shapes how he thinks about today’s AI: not as algorithms in isolation, but as architectures whose behavior, like the brain’s, emerges from structure. After 32 years as a professor at SMU, where he taught across nearly every corner of computer science, he retired to focus on the technologies reshaping the field.

He now teaches generative AI and large language models at the University of California, Berkeley, and holds a visiting professorship at the University of Bologna’s graduate school of business. He is also the founder of Edge AI (codesupreme.ai), where he develops courses and tools that help working developers build reliable, production-grade agentic systems. Known for making difficult technical ideas clear and memorable, Frank teaches audiences ranging from Berkeley graduate students to district attorneys and formerly incarcerated learners — and, most recently, the 250,000+ viewers of his AI Engineer World’s Fair talk on why agentic systems need ontologies.

Talk details

Talk
AI/ML Intermediate

Why Agentic Systems Need Guardrails: The Case for Ontologies

Frank Coyle, PhD
Frank Coyle, PhD Lecturer @ UC Berkeley | Visiting Lecturer, University of Bologna
Agentic Engineering AI Architecture governance

Agentic systems fail in predictable ways. A second refund on the same order. A payout sent to the support desk instead of the buyer. An order status of “probably shipped.” These aren’t random glitches — they’re symptoms of one missing layer.

LLMs reason probabilistically over domains they only partially understand, and no amount of prompt engineering fully closes that gap. This talk argues that the missing layer is an explicit ontology: a formal, shared map of a domain’s concepts, relationships, and constraints, sitting outside the model as enforceable logic.

The result is a hybrid neurosymbolic architecture — probabilistic reasoning inside, logical guardrails outside. Drawing on a pattern that has quietly powered enterprise systems for over a decade, the session shows how lightweight ontology constructs surround an agentic system with rules it cannot violate — turning brittle, unpredictable agents into ones you can actually trust in production.

Read more →

Data

Leadership

Product

AI/ML

UX/UI

Join the event!

Stand on the shoulders of giants and build alongside the people shaping what comes next.

Save the Date

Add the 2026 conference to your calendar