I am building a Resilience Intelligence Layer (RIL) - the OS for infrastructure resilience. RIL is an AI-enhanced epistemic framework for earthquake-resilient infrastructure systems. The goal is to build AI systems on high-fidelity physics-grounded models that create, test, and refine epistemic gaps of infrastructure resilience under large seismic events so decision-makers get robust, interpretable, and actionable resilience prescriptions. Current earthquake hazard and risk models either (a) simulate physics at scale (expensive), or (b) learn correlations from data (non-explanatory). We need a hybrid, explanatory scientific system that proposes hypotheses about cascading failure mechanisms from ground deformation to infrastructure systems failure, uses physics to test them, iteratively refines knowledge, and enables faster scientific insight and resilience decision-making.

Previously I evaluated statistical variability of building damage response by utilizing a suite of HPC earthquake hazard simulations for the San Francisco Bay Area. This knowledge is critical for engineers to design new buildings and retrofit the existing structures. Furthermore, insurance companies require this information to account in their seismic risk evaluations based on geographical and geological parameter differences. This foundational research has enabled me to embark upon building the current RIL framework.

I am also exploring and developing research software capabilities for floating seismic isolation systems. This project is aimed at enhancing seismic safety of nuclear power plants by modeling fluid-structure interaction and computing the fluid-induced seismic isolation effects on the structure from both horizontal and vertical earthquake motions.