This time Google 26NG SDE VO After three rounds of interviews, the overall pace was relatively stable, but there were still many pitfalls in the details. Many people think that Google interviews are just about studying algorithms. In fact, the real experience is more like examining communication, ideas, and code details. The core process and question sharing of each round have been organized for your reference. Student background: Not a top school, 3-4 years of experience. Previously worked as a back-end in a large factory. LeetCode has completed about 800-1000 questions.
Round 1: Pure Behavioral
The interviewer for this round was an Asian guy, who was really nice. At the beginning, he took the initiative to chat with me for seven or eight minutes, talking about the weather, what dramas he was watching recently, how many years he had been in the United States, and other topics. There was no pressure at all like an interview. After briefly introducing himself, he went directly to three standard behavioral questions.
The first one is for me to share a practical case on how I identified and resolved a technical risk. I talked about the potential SLA instability caused by relying on third-party services in previous projects. How we discovered the problem in advance through monitoring and stress testing, and designed a downgrade plan, and what the effect was after it was finally launched.
The second one describes how I sorted out task priorities and ensured on-time launch when faced with a particularly tight delivery deadline. I used a method similar to RICE scoring combined with MVP splitting and daily quick stand-up meetings to align the entire team.
The third part is to share an experience of receiving critical feedback on the code. I talked about how a senior pointed out the problem directly in the PR. At that time, I accepted it humbly and then made more systematic improvements, including improving code readability, test coverage and annotation specifications.
Round 2: BQ + Coding
The second-round interviewer was an Indian engineer. His English had a distinct accent and his speaking speed was relatively fast, but the logic of the questions was actually very clear. I asked two BQ questions first, and then took two coding questions directly in succession.
BQ part
Q1. Why do you want to work at Google? Can you share your thoughts?
Q2. Is there any project that you feel you have done particularly well and exceeded everyone’s expectations? Specifically?
Coding
The first is the classic conversion of a binary search tree into a sorted circular doubly linked list. I used in-order traversal, maintaining a tail pointer, establishing a two-way connection while traversing, and finally connecting the head and tail to form a ring. Pay special attention to the boundaries of returning null for an empty tree, connecting a single node to itself to form a ring, and finally returning the smallest node as the head.
The second question is to query an array and multiple intervals, requiring to determine whether each sub-array satisfies the odd-even alternation of adjacent elements. My approach is to preprocess an auxiliary array first, record whether the parity of two adjacent numbers is different, and then transform the problem into judging whether the auxiliary array in the interval is all 1. The prefix sum can be used to quickly answer, and the overall time complexity is O(n + q). The interviewer agrees with the idea of preprocessing.
This round was a bit tiring due to the accent and the two consecutive coding tasks, but overall it went smoothly.
Round 3: Pure Coding
The interviewer in the third round was an ABC. The atmosphere in the whole round was the most relaxed. I hardly asked behavioral questions. I started coding directly after a few greetings at the beginning. He was very kind and would take the initiative to ask me about my ideas. He would also give more natural tips when I got stuck.
The first question is to give a binary tree and ask to return the node values that can be seen from the bottom up when viewed from the left, and to return the node values that can be seen from the top down when viewed from the right. I used a layer-sequential traversal method. Each layer records the leftmost node and the rightmost node. The results of the left perspective need to be reversed at the end. The right perspective is retained directly from top to bottom. Note that the root node will not appear repeatedly on both sides.
The second step is to find the longest strictly increasing path length in the binary search tree. The path can start and end from any node. I used DFS recursive writing. Each node returns the longest downward increasing length starting from itself, and uses global variables to update the maximum value of the left subtree chain + right subtree chain + 1. Because it is strictly incremental, although it is BST, it cannot actually take advantage of the size relationship between the left and right subtrees, and it can be passed in O(n) without pruning. After I finished writing, the interviewer asked if I could use the properties of BST to optimize it. I simply said that I could add some judgment but the benefits would not be big, so he nodded and approved.
This round was the most comfortable to write, and the interviewer’s guidance and atmosphere were particularly good.
Learn more
Many candidates will encounter situations where time is extremely tight, question variations are endless, and psychological pressure is high. At this time, if you can find a truly reliable professional auxiliary service, you can pass these key levels more efficiently. If you are anxious about the next round of interviews, or feel that you don’t have enough time or confidence, please contact us in advance. Contact us Let’s talk about your specific situation, and I will personally help you evaluate the most suitable plan, so that your job search will be less detoured and more confident.
