Experiment : hire 10% of candidates without conducting any interviews. Use Machine Learning to quantify internal or external candidates' code quality. Use that

Experiment : hire 10% of candidates without conducting any interviews. Use Machine Learning to quantify internal or external candidates' code quality. Use that model to score candidates' GitHub. Hire them. Well, OK, throw in a sociability / not-a-psycho score using LinkedIn or twitter or whatever. I'd outsource this to gild.com who seem to know what they're doing. Methinks management will be surprised at how well these 10% do despite not passing the usual "What happens when I flip all the color bits in a red-black tree" or whatever shibboleth is being used these days. Admit it : most work, even at Google, doesn't involve knowing how self-balancing trees work. On the other hand, getting things done matters everywhere.