How would you clean and integrate messy, heterogeneous pharmaceutical data from multiple sources at Servier?
