Status : SDE (ML)\nRelevant YOE : 2\nEducation : Masters\n\n**Google**\nPosition : SWE - ML\nLocation : Pittsburgh\n*Referral from a former colleague*\n\n**Phone Screen 1**\nTime : 45 minutes\nQuestion 1 : Given an array find which numbers occur fewer than K times in it.\nQuestion 2 : [Hand of straights](https://leetcode.com/problems/hand-of-straights/)\n\n**Phone Screen 2**\nTime : 45 minutes\nQuestion 1 : [Median in a Data Stream](https://leetcode.com/problems/find-median-from-data-stream)\nQuestion 2 : Given an excel spreadsheet and a formula/constant in each cell, find if there is any such cell whose value cannot be computed. For example,\n**Cell = Content**\nA1 = B1 + 2\nB1 = C1 + 4\nC1 = A1 + 6\n*Essentially, detect a cycle in a graph. Use any method you\'d like.*\n\n\n**Virtual Onsite**\nTime : 4 hours ( one 15 minute break )\n\n**Interview 1** :\nBehavioural - very similar to Amazon\'s leadership principles\n\n**Interview 2** : \n1. [Inorder Successor in Binary Search Tree](https://leetcode.com/problems/inorder-successor-in-bst)\n2. Given a set of points as centers of circles with radius r and a 2D grid of size WxH find which sets of circles overlap and return them as groups\n\n**Interview 3** :\nGiven a source string ("abelphata") that was generated using a merge() function and a dictionary object that has only one method called contains() return the original two strings ("alpha" and "beta") that generated the source string.\n\nConstraints:\n* You must treat the merge() function & the dictionary as black-boxes. \n* You may also not iterate through the dicionary.\n\n**Interview 4** :\n[Guess the Word](https://leetcode.com/problems/guess-the-word)\n**Modification** - the master unlocks the characters and their respective positions which are correct from your guess.\nFor example, if the target word is "tiger" and you guess "ridge", the master returns "_ i g e r"\n\nConstraints :\n* You can only guess from the dictionary\n* You start with $10,000 and for every incorrect guess you lost $1,000. After the first guess for every millisecond elapsed you lose $10.\n* You must maximize your chance of getting the right answer on the first guess\n\nNote - Owing to the third constraint some of the leetcode accepted answers such as synthetic guesses, pick a random word, pick the first word etc. aren\'t acceptable. The interviewer was adamant that there\'s a better way and it took me 15 minutes to find it and code out.\n\n**Interview 5** :\nML and Computer Vision specific questions regarding how I would design classifiers, object detectors etc to be used in random Google products.\n\n

Status : SDE (ML)\nRelevant YOE : 2\nEducation : Masters\n\nGoogle\nPosition : SWE - ML\nLocation : Pittsburgh\nReferral from a former colleague\n\nPhone Screen