Timeline: mass-applied -> OA -> HR screen -> Round 1 -> Round 2
The HR round just went over my resume, timeline, and identity/status information. None of the brain teasers people mentioned in earlier interview experience posts.
Round 1 was an open-ended ML system design.
The question: after a basketball is thrown, it can either go in or miss — how do you model this using information available at the moment of the shot to predict the outcome? First I confirmed this is a classification problem. Then we discussed the data: basic features like the ball's release speed and the distance to the hoop, a player embedding (could use an nn embedding), and contextual info like the temperature that day and how long the player had already been on the court. After that we discussed how to build the model and what loss to use. Then we talked about feature selection — an interesting question was: if you run a linear regression and regress y on a single feature by itself and it comes out not significant, can you drop that feature?
Round 2 was coding.
It was the Connect Four game that's come up in other people's interview experience posts — I solved it quickly. Given a 6x7 board, two players alternate choosing a column to drop a ball into; because of gravity the ball falls to the lowest empty spot in that column. If a player gets four in a row — horizontally, vertically, or diagonally — they win. Then there were follow-ups: what if there are more than two players, and what if the board is very large?
Discussion
Loading comments…