Organic-inorganic hybrid perovskites have emerged as promising light absorbers in photovoltaic (PV) cells or as emitters in light-emitting diodes (LEDs) with the goal of achieving high performance using various device structures, fabrication methods, and new doping or substitution of ionic components. These new perovskites allow integration of useful organic and inorganic characteristics within a single crystalline molecular-scale hybrid, enabling unique electronic, magnetic, and optical properties. However, the surface stability and interfacial compatibility of hybrid perovskites, which restrict their practical applications, have been largely unexplored at the atomistic level. Computational and dataenabled materials design can indeed shed light on experimental observations and predict materials properties which are difficult to access experimentally. The proposed research combines molecular dynamics (MD) simulations, Monte Carlo (MC) simulations, force field development from quantum mechanical (QM) calculations and existing experimental data, and Bayesian statistics-based uncertainty quantifications. This will yield a predictive modeling and design framework for hybrid perovskites at larger time and length scales. The synergistic combination of potential of mean force (PMF) calculations with surface wetting, vapor adsorption, and reaction kinetic theories can lead to accurate perovskite lifetime predictions. With this fundamental understanding in hand, we will design chemically-stable hybrid perovskites passivated by water- and vapor-resistive ligands.