Twitter is a global, high-traffic operation with a pile of dedicated servers. It takes a lot of facilities staff to deal with that.
Twitter is an HA service that is trying to build a reputation for reliability. You need enough ops staff to keep a 24/7 NOC up and running and to deal with illnesses and vacations.
Twitter is in an explosive growth phase. They are quite likely to be employing lots of engineers on strategic projects. It can be a mistake to think of your R&D team as some kind of cost that you need to contain. R&D is an investment in the future. And Twitter has a lot of VC money to invest right now.
Not everything scales. In particular: Sales and customer service don't scale perfectly unless you deliberately engineer that from day one. For example, if you build Google Ads (and are able to compete with the original Google Ads) you can sell ads with Perl scripts. But traditional ad sales requires salespeople. Customers like to be talked to, they like to get their problems solved and their tickets closed by sympathetic humans. Depending on the various business models that Twitter is considering, they may need a bunch of sales people to negotiate deals with customers, and/or a bunch of systems integrators to help customers learn to process the data coming out of the firehose.
Twitter needs lawyers and moderators and ombudsmen. They receive DMCA takedown notices. They receive reports of threats and spam. People steal credentials and impersonate other people. There are privacy laws and telecom regulations. They have to deal with these things in multiple languages and around the world.