來這裡累積個人品，面經如下：\n\nPart 1: Python coding .\n烏伯高頻題，求 sqrt 那個，用 Newton's method 秒殺，考官有提到大部分人用 binary search，總之有準備就不會太難。\n\n\n\nPart 2: Case study\nHow to decide which of the new and existing machine learning models is better at determining which users receive a $5 promo?. Χ\n主要是在問success metric, model selection, and evaluation of model performance，還問了 a/b testing細節，然後有問一點點 causal inference 概念。\n. ----\n下禮拜要面 Jam session 希望會過...