Girish Patangay ,Engineer at Facebook.
I don't think you can have "usual" and "scaling issues" next to each other. Scaling at our size pushes every piece of infrastructure we have to it's limits and beyond.
Besides the 'usual' work I'd say there is a lot of work being done in making things more efficient. For example, we are working on the JIT for HipHop.  Something completely at the other end of the spectrum is the work that is going on in building our own custom data centers that are highly efficient along with our own custom hardware. 
I, for one, have been focused on making our engineers more efficient. I've been approaching this very subjective problem from multiple angles. Making it easier for engineers to get their work done, and after they are done with their work, making it easy and foolproof to verify and test their changes.
Check out our engineering blog posts to see a lot of other things we've been working on!
0 blog comments below