Given a large dataset at Eventual, how would you optimize a specific transformation to minimize CPU and memory overhead?
