I think there are multiple different type of A player developers: There are "researcher" developers that given a tough problem can come up with an excellent so

I think there are multiple different type of A player developers: There are "researcher" developers that given a tough problem can come up with an excellent solution. If you gave the same problem to someone less competent, you would get a hack with lots of problems, or possibly no solution at all. There are also "coder" developers that given a well fleshed out specification, can crank out code at break-neck speed. You can tell research developers in interviews by proposing large-scale problems and asking them how they'd approach finding solutions. Do they get excited, or freeze up? Big tell. To get coder developers, open source project activity is probably the best indicator. If someone is part of several open source project teams and has active forks and a commit history, that is a good indicator. Once you have people like this, when managing them, you need to make sure that you assign tasks based on their strength. Coders are great as project maintainers, researchers will probably find a new job if you put them in a maintainer role. Likewise, coders aren't going to be happy if you don't have well fleshed out requirements/spec for them to chew on, while researchers view that as a license to let their creativity run wild.