Clouds have been extensively studied, yet they remain one of the largest sources of uncertainty in contemporary Earth System Models (ESMs). This project aims to reduce these uncertainties by integrating ESM output with observational data. In the first part of the work, observational constraints on cloud feedbacks in response to climate change will be investigated, with the goal of narrowing their spread and improving confidence in model projections. The second part focusses on another aspect of clouds, looking at aerosol-cloud-interactions and their importance for global climate, weather forecasting, and precipitation patterns. For this, new methods to combine recent advances in machine learning weather prediction (MLWP) and new high-resolution observational and numerical modeling data will be developed.