Sarah Mokhtar is a researcher and architect working at the intersection of machine learning, building science and computational design.

She recently defended her PhD dissertation in Building Technology at the Massachusetts Institute of Technology (MIT), where her research developed expressive representations and simulation-driven learning frameworks for the built environment. Grounded in building physics and environmental modeling, her work advances scalable approaches to reasoning about geometry, climate and performance across buildings and urban systems. Her research integrates machine learning with physically informed modeling to support climate-resilient, performance-aware and human-centered design.

Prior to MIT, Sarah worked as a Senior Environmental Performance and Computational Designer at Kohn Pedersen Fox Associates (KPF). She led applied research projects in the fields of machine learning, simulation and computation and managed KPF's Environmental Performance (KPFep) team in the London office. Her work focused on developing computational tools and physics-based workflows to support evidence-based and performance-driven design.

Her research has been published in leading journals including Building and Environment and Automation in Construction, and presented at major international venues spanning building simulation, urban climate, computational geometry and machine learning, including the International Conference on Urban Climate (ICUC), Advances in Architectural Geometry (AAG), Symposium on Simulation for Architecture and Urban Design (SimAUD), Asian Conference on Machine Learning (ACML), IBPSA's Building Simulation Conference and NVIDIA’s GPU Technology Conference (GTC). Her work has also been showcased at the 2025 Venice Biennale.

She holds an MSc in Adaptive Architecture and Computation with Distinction from The Bartlett School of Architecture, University College London (2016), and a BSc in Architectural Engineering with Highest Honors from the American University in Cairo (2014).