We have a PhD computer scientist working as as software engineer. Reinvents the wheel constantly. Over engineers everything. Clever algorithms but poor abstract

We have a PhD computer scientist working as as software engineer. Reinvents the wheel constantly. Over engineers everything. Clever algorithms but poor abstractions. Needs to have the "best" (from a CS perspective) solution to everything rather than the most maintainable or pragmatic solution. He is smart, but doesn't write code that is easy for the next developer to pick up.