[bradcomp]

The biggest thing I see is not considering edge cases.

They'll have a certain way they are checking a few cases (and maybe a couple tests), but they don't have the foresight to really try to break things. They will also make assumptions about how the code will be called, what entities will exist at that point, etc. that just don't hold in reality.

This typically manifests in code that works fine when they use it, and all the tests pass, but it breaks as soon as it gets up on the staging server, or as soon as somebody else tries to use it.

Also: Not listening to feedback from senior devs. Drastically underestimating the complexity of what they are working on.