A decade ago, buildings and infrastructure systems began to transform by becoming more instrumented, interconnected, and capable of generating immense streams of data. Yet despite this progress, their operations remained largely reactive, guided by static rules and limited models.

Farrokh Jazizadeh, associate professor, envisions a different future - one where our built environments don’t just respond, but anticipate by learning from context, adapting to change, and optimizing for human experience. His research sits at the crossroads of AI, human-technology interfaces, and smart infrastructure, exploring how generative AI could unlock this shift and redefine our relationship with the spaces around us.

Recent advances have opened a door. Large language models (LLMs) and foundation models, trained on vast and diverse datasets, now offer the reasoning

power to interpret multimodal data streams, support natural interactions, and generate adaptive solutions. This scalability could remove barriers that once held back AI-powered infrastructure and set the stage for a new era of operational intelligence.

In his lab, Jazizadeh’s team is building toward this vision. One effort develops intelligent ambient AI agents that combine contextual reasoning with human-in-the-loop decision-making. Another advances scalable foundation models for digital twinning and predictive operations, creating systems that adapt seamlessly to new environments while outperforming conventional machine learning approaches.

The story of tomorrow’s infrastructure is no longer one of rigid systems. It is one of intelligent partners that understand, predict, and evolve alongside us, shaping environments that feel as responsive and dynamic as the people who inhabit them.