A Carnegie Mellon team used an automated system and machine-learning software to develop fast-charging electrolytes that outperformed a standard one.