Laurence Moroney

Most organizations have run AI pilots. Far fewer have successfully moved those pilots into production at scale. Moroney draws on his experience building AI infrastructure at Google and now at Arm to give audiences a frank, technical, and immediately practical account of what separates proofs of concept from production systems. He covers model selection and fine-tuning, the infrastructure decisions that determine whether AI compounds value or creates overhead, the governance and evaluation frameworks that keep deployed models reliable, and the organizational patterns he has seen succeed and fail across thousands of teams worldwide. A keynote for technical leaders, product managers, and executives who need to move beyond the experimentation phase.

The dominant conversation about AI assumes cloud-based inference on large models. But the next major shift in AI deployment is happening at the edge: on smartphones, vehicles, industrial sensors, and the billions of Arm-powered devices that make up most of the world's computing infrastructure. In this keynote, Moroney presents the technical and strategic case for on-device AI, covering the current state of model compression and quantization, the privacy and latency advantages that make edge inference compelling for specific use cases, and the developer tooling that is rapidly closing the gap between what is possible in the cloud and what is deployable on device. A forward-looking session for organizations in manufacturing, healthcare, consumer technology, and any sector where real-time inference matters.

As AI becomes part of how people learn, work, and make decisions, the question of whose language, values, and context a model reflects becomes a strategic one. Moroney explores why culturally specific models are growing in importance, how smaller models running on device can protect privacy and preserve local knowledge, and what organizations operating across markets should weigh when choosing between global foundation models and tailored ones. A timely session for multinational companies, governments, and any organization serving diverse audiences.

For technical audiences and developer-heavy organizations, Moroney offers a deep-dive session based on his bestselling O'Reilly books, including AI and ML for Coders in PyTorch, and the PyTorch for Deep Learning Professional Certificate he teaches with DeepLearning.AI. He walks teams through the practical architecture of modern machine learning systems, covering data pipelines, model training and evaluation, deployment patterns, and the specific trade-offs between TensorFlow and PyTorch for different production contexts. The session is calibrated to the technical level of the audience and can be delivered as a keynote, a workshop, or a hybrid format combining strategic framing with hands-on exercises.

One of the defining challenges of the current AI transition is not technical but organizational: most companies have concentrated their AI capability in a small number of specialists, while the rest of the organization watches from the outside. Moroney has spent his career arguing that this model is both unnecessary and limiting. Drawing on his experience designing AI education programs for millions of learners, he presents a practical framework for building AI literacy across functions, levels, and technical backgrounds, sharing what works in terms of curriculum design, tooling choices, and the cultural conditions that determine whether AI capability compounds or stagnates inside an organization.