In a machine learning project that requires processing a vast amount of data spread across multiple nodes, which Spark ML component is optimized for efficient distributed computing and handling of large-scale datasets?