When you are at the cutting edge of your field (able to publish in Nature or Science), you know exactly who the peer reviewers are (there is usually only 5 to 1

When you are at the cutting edge of your field (able to publish in Nature or Science), you know exactly who the peer reviewers are (there is usually only 5 to 10 teams in the world who are experts in your area). So you know it's one of 20 to 60 people (roughly). That part is actually very easy. The job the journal does is to randomly and anonymously assign the reviewer (sometimes not entirely randomly, giving it to your critics). But all that could be automated (taking into account top publishers in the field, newcomers to the field, conference publications, etc).