The first step is to have a good coder who can organize their commits into simple self-contained small steps that can be reviewed individually. Then your job ge

The first step is to have a good coder who can organize their commits into simple self-contained small steps that can be reviewed individually. Then your job gets 10000% easier. If I giant mess of a commit ends up assigned to me I just send it back or stall. If you want my attention you have to help me help you. The Linux Kernel communities have a lot of good advice on both writing self-contained patches and properly reviewing things. As an extra advice, code reviews can often feel very personal and are a sensitive topic. Always make sure to be impersonal: instead of "you did this and that, which is not correct" try to go with "the code does this and that, which is not correct", and "we should do that" instead of "you should do that".