I wonder if an interview process more like academic science would help. It usually goes like this: Candidates are roughly screened by their CV. A handful (in

I wonder if an interview process more like academic science would help. It usually goes like this: Candidates are roughly screened by their CV. A handful (in my experience roughly 6) are invited to hang out at the lab for most of the day. They usually give a presentation on previous projects, and then chat with each member of the lab in turn, hearing about their work and asking questions etc. Then you all go and have lunch together (usually without the boss). Later the candidate has a more formal panel interview with the group leader and some other faculty. A few days later, they are told the outcome of the interview. It may seem like a long process, but it all happens in one hit. The advantage is you get a much better feel of the candidate over this continuous process, than you would with shorter interview tasks.