SARAH MOKHTAR

Sarah develops data-driven machine learning frameworks for reasoning, synthesis and decision-making in complex physical environments. Her work explores spatial intelligence to support climate-resilient, human-centered and performance-aware design across scales.

Sarah’s research is driven by a central question: how can learning systems reason about and improve built environments in a rapidly evolving world with shifting environmental, social and performance demands? She develops computational and machine learning frameworks that connect geometry, environment and performance, enabling informed and adaptive design and decision-making across buildings and cities. By grounding ML in real-world physical systems and their structural constraints, her work translates heterogeneous spatial data into actionable representations. As both an architect and machine learning researcher, she integrates domain structure, spatial intuition and environmental rigor into representation and modeling approaches grounded in real-world systems and constraints.