Why don’t we apply the collaborative filtering approach to finding the right job candidate?

Here’s a simple model of how it could work for say, a network engineer:

Any and every potential candidate is invited to submit potential questions to ask which they think could seperate out a good network engineer from a bad network engineer over the course of a 24 hour period.

Once these questions are accumulated, all candidates are split into two groups and given one hour to use a collaborative voting system to determine which questions they feel are the best ones.

Each group gives the top n questions to the other group and they both have 3 hours to complete the test.

Each group now collaboratively marks the other group. A right answer is one which concurs with the answers of those who got the most right answers. In the end, the top 5 people with the highest score from each group are selected for a in depth interview.

Is this approach better than the typical HR keyword search based weeding approach? Is it robust enough to efficiently weed out the poor candidates while pushing the good ones to be great? It seems like an interesting experiment to me.

Leave a Response