The author’s Voleon phone-screen report combines mathematical reasoning, statistical modeling, and a short programming task. Matrix questions concerned the implications of unit row and column totals, including whether nilpotence was possible. The regression discussion examined bounds on the coefficient of determination and what could happen when two predictors were used together.
Further questions tested how to interpret training and test errors, with attention to the effects of regularization and dataset size. For coding, the interviewer supplied a Word document containing a task about recognizing words with repeated letter-frequency patterns. The report lists the topics without giving the author’s answers, interviewer feedback, or hiring result.
Discussion
Loading comments…