I never work for Uber but I can easily imagine the following problems:
* Map data is not a solved problem. Tons of inaccurate data especially on non-english speaking countries.
* As someone mentioned about payments, different country requires different payment platform integrations.
* Payment is just one aspect of money, you need to build a treasure-like logic to run a marketplace between drivers and passengers.
* The matchmaking algorithm itself is interesting enough problem. It has to be efficient for the company to make the most amount of money.
* Tons of internal data are produced and required to be understood. This requires ML/data analytic talents.