The Material Vision team is developing a computer vision algorithm, 'Cosmos,' to optimize thin-film halide perovskite solar cells by achieving uniform film coverage, which is crucial for enhancing performance and long-term durability. Cosmos collaborates with Dr. Carolin Sutter-Fella’s team at LBNL to analyze photoluminescence images, generate a heterogeneity score, and provide suggestions for improving the manufacturing process to lower this score. By following these insights, the goal is to produce a more uniform film with better optoelectronic performance.