Amazon Interview Questions

Scott S. - "To recount an experience from my last project . We were at the initial stages of designing a new data platform to enhance our trading analytics capabilities. The conflict arose over the selection of a data processing framework. I was advocating for Apache Spark due to its scalability and performance benefits, especially for handling large volumes of derivatives data. Another senior team member preferred a different technology they had more experience with, which they believed would be easier to"See full answer

Tes d H. - "Here is my first shot at it. Please excuse formatting. To find the maximum depth of the dependencies given a list of nodes, each having a unique string id and a list of subnodes it depends on, you can perform a depth-first search (DFS) to traverse the dependency graph. Here's how you can implement this: Represent the nodes and their dependencies using a dictionary. Perform a DFS on each node to find the maximum depth of the dependencies. Keep track of the maximum depth encountered dur"See full answer

Pm P. - "Alexa: Is the virtual assistant that runs on amazon smart speakers as well as on other devices. Clarification on goals/Objectives: What would the goal at that point of time be? Is it increase in revenue ? Either via new customers i.e. adoption or Increasing the wallet share of the customer? Increasing the sales ordered via Alexa ? Why do people want to use Alexa? To get answers, make their lives easier by automating activities within their home Integrates with amazon account so"See full answer

Adib M. - "Useful article on API migration https://www.linkedin.com/pulse/tips-successful-api-migration-pooja-nihalani/"See full answer

Anjali M. - "Colleague moved to a different role, so I decided to fill in and did outshine in showing results within the group."See full answer

Surbhi G. - "While serverless computing has its merits due to removing the need for server management and auto-scaling abilities, several challenges might arise when deploying machine learning models using this setup. Cold-start latency is a key issue where an Amazon Lambda function, for example, would take more time than usual to run after being idle. This could affect the performance of real-time ML applications negatively. A way to keep the functions warm might be required to avoid this latency. Moreover,"See full answer

Parag L. - "I was in my early days of leadership and strategy. In Marriott when a decision was made by C suite to buy products as wrapper over cloud providers I was surprised and based on my limited information and knowledge posted a question to Chief technology officer as why it won't be a good strategy to directly use Native offering from AWS. The reason enlightened me and gave new a new thought process that being such a large enterprise that has presence in 200+ countries across globe we may have challen"See full answer

Ben G. - "In my time as a product designer, I’ve seen a transition of growth from something that is visually appealing to what is actually useful / functional / scalable. I think as designers, we are afraid to let things go such as the visual fidelity of design and how something could be useful but not necessarily have the flash of a more consumer based product. I think when we learn to let that go, we can see the scale of our impact and that we aren’t just people who make something look good/pretty but t"See full answer

Amparo L. - "My parents because they take really good care of me whenever I feel sick"See full answer

Khushbu R. - "Let’s say the matrix is m x n (i.e., m rows and n columns). Start from the top-right corner of the matrix. Move left if you see a 1. Move down if you see a 0. Keep track of the row index where you last saw the leftmost 1 — that row has the most 1s. public class MaxOnesRow { public static int rowWithMostOnes(int matrix) { int rows = matrix.length; int cols = matrix[0].length; int maxRowIndex = -1; int j = cols - 1; /"See full answer

Suhasan C. - "Clarifying questions: In which environment are we using the AmazonEcho? in mobile devices? Car? Echo devise at home? I think the following groups make the best target according to the order If the Echo is in-home devices: Kids and Teens Elderly Physically Challenged Eg. Blind or Deaf Others If the Echo is in a Car/Vehicle: Uber/Lyft or Taxi drivers Long-distance travelers taking road trip"See full answer

Ben G. - "Situation - A time I saw a peer struggling was while I was at Google working in Mountain View. A fellow colleague of mine was consistently not meeting standards in his performance reviews and wasn’t sure why. He was upset because Google was his dream company and he wasn’t sure how to improve based on his feedback. Task - I was tasked to design a based VUI framework for our discovery space on smart displays. This colleague was acting as the VUI lead for the project while I acted as t"See full answer

Ds S. - "Initially I asked clarifying questions like whether the tree can be empty or not and asked the interviewer to explain what is meant by left view and the explanation for the sample inputs. Then I came up with the level order traversal approach where we visit each level in the binary tree at once using a queue and at each level print the value of the first node. Interviewer seemed satisfied with the approach and asked me to code it up. Finally gave the time and space complexity of the solution."See full answer

Jaim - "Whenever I find resistance to my idea, my general approach is to : Validate my idea Identify if it is a political reason or is it a valid engineering tech reason Assuming it is not political but it is resource/difference of opinions that are causing resistance. Then I would try to do a quick POC and show a value proposition for the idea to get the buy-in. One similar situation happened The situation was that stakeholders were promised a feature that would decrease their number of ticket"See full answer