Amazon Applied Science On-Campus Internship Interview Experience 2020
Amazon recently came(virtual) to our college for intern hiring for the applied scientist position. The students were shortlisted based on a resume from CSE/MnC. There were two rounds, first was on coding, and the second was on resume projects and core machine-learning algorithms usage and implementations(pretty hard).
As usual, Amazon has arranged it is an online interview through amazon chime. Here is my experience….
Round 1(60–70 minutes):
It was a coding round, and the interviewer was polite and had good knowledge of competitive programming. The interview started with an introduction, and he talked about the current situation(pandemic) and some personal information and made me comfortable for the interview.
There were six coding questions. The interviewer asked me to write the optimized code for three questions.
1- Given an array, find the maximum possible sum such that no consecutive elements should be added in the sum. (He asked me to write the code of this question in the text editor)
2- Given a binary tree, find the sum from the root node to any given node.
3- Given a tree, find the sum of all nodes between given two nodes(both inclusive).
4- Given a 2-D matrix, contains char “.” (can move), “#” (blocked cell), “P” (Starting position of person), “E” (Ending position). The person can move in L, R, U, D directions(if possible), print the 2-D matrix with minimum move path from P to E and represent the path with “*” in the 2-D matrix.
5- Given a 2-D matrix, contains char “.” (can move), “#” (blocked cell), “P” (Multiple starting positions of person), “E” (Multiple ending position). The person can move in L, R, U, D directions(if possible), print the 2-D matrix with minimum move path from every P to the nearest E and represent the path with “*” in the 2-D matrix. (He asked me to write the code in the text editor)
6- Given an integer array, print the randomly permuted array so that no one can find the pattern of randomness. (He asked me to write the code in the text editor)
In this round, I answered all the questions, and the interviewer was very satisfied with my answers. No. of the questions was not fixed and depend upon how many problems you can solve in 60–70 minutes.
Round 2(60–70 minutes):
It was a machine learning, data-science round, and the interviewer was polite and had pro-level knowledge of machine-learning algorithms. The interview started with an introduction and went into an in-depth discussion of resume projects and four machine-learning algorithms.
Also, the interviewer had a habit of writing the notes from the discussion. He discussed the resume project and ML algorithms, and then at the end of the discussion, he was writing the points from the discussion and confirmed the same from me.
1- Tell me your best machine learning projects. Then he discussed very minute details of the project. He asked me why this project and how this is helping the community. Why did you do in this way? If there are other methods, why did you do the same project again? Did your project algorithm beat different algorithms in terms of performance and computation power? (Around 20 minutes discussion)
2- Are you comfortable in machine learning algorithms? Do you know Logistic Regression? What is Logistic Regression? And What is the principle behind the working of this algorithm? (Around 20 minutes discussion)
3- What is the k-means algorithm? Can you tell me the working principle of k-means algorithms? Let’s say if I will ask you to make a k-means algorithm for me, then what are the things you will need, or you will ask from me? (tense discussion) What are the distance metrics you will use to measure the distance from the cluster center and points? What are the different distance metrics you know? Why will you use this particular distance metric? When will you use one over another one? (Around 20 minutes discussion)
4- What is the ensemble? What is the bagging and boosting algorithm? Describe the working principle of both algorithms. When we will have high variance algorithms, what will you use bagging or boosting? (Around 10 minutes discussion)
In this round, “describe the working principle” needed profound knowledge of the algorithm and you can’t fool the interviewer or say anything. So for Amazon Applied science(AS), you need to have a profound understanding of the machine-learning algorithms. Also, I had given most of the answers, and I was confident, but Amazon hadn’t taken anyone for the internship role from the campus. Yes, you heard it right.
