I'm a post-doc at a university. I asked around at conferences, and emailed people whose work I liked. I'd be surprised if the process is very HPC-specific; you'd be trying to make scientific job-hunting more efficient. But please do that if you can :D