<p>Perovskite types of photovoltaic sources have become the primary candidates for next-generation photovoltaic sources thanks to their high-quality optoelectronic characteristics and their fast-enhancing power-conversion parameters. Nonetheless, the computational space of perovskites is quite large and first-principles is expensive, and this makes systematic optimization of materials a challenging problem. This paper presents a unified multiscale system that entails the implementation of density functional theory (DFT), machine learning (ML), and device-level simulation to hasten the search and development of high-performance perovskite solar cell materials. Physically significant descriptors are generated through DFT calculations such as structural parameters, electronic bandgaps, formation energies, and properties of defect-related characteristics. The descriptors are utilized to train the supervised ML models that can predict the important material and photovoltaic performance indicators with high accuracy at a rapid rate. Predictions of the ML are then combined with practicing simulations of the device to assess open-circuit voltage, short-circuit current density, fill factor, and power conversion efficiency. The findings show that there are good correlations among defect tolerance, bandgap optimization, and photovoltaic performance with defect suppression being a leading channel of increasing the efficiency. The current DFT-ML scheme consumes much less computational power; however, still has physical interpretability, and can offer efficient and predictive solution to perovskite material screening as well as solar cell design scaling.</p>