Amazon

I applied for the Applied Scientist Intern opportunity at Amazon through the Amazon ML Summer School route. The entire interview process took around two to three weeks. Online Assessment The first stage was a HackerRank online assessment. It had two easy-level DSA questions. Compared with the usual Amazon SDE assessments, the coding questions were relatively straightforward. The main focus was on getting the implementation correct within the given time. I was able to solve both questions and subsequently received the interview invitation. Technical / DSA Round The first interview was around one hour. The interviewer started with a few introductory questions and asked about one of my projects before moving into DSA. The first coding problem was a shortest-path problem in a graph with unit edge weights. I explained the BFS approach, including why BFS guarantees the shortest path when every edge has the same weight. The second problem was based on a Binary Search Tree. The interviewer wanted me to reason through the implementation rather than simply use a library implementation. I explained the recursive approach and discussed the complexity. The interviewer also asked several follow-up questions about edge cases and why the selected data structure was appropriate. The round was mainly focused on problem solving, with some discussion around my projects. Machine Learning Round The second round was conducted the following day and was completely focused on Machine Learning. The first part of the interview was based on my ML projects and previous work. I had to explain what problem I was solving, how I prepared the data and why I selected particular models. The discussion then moved into Machine Learning and Deep Learning fundamentals. Questions covered topics such as RNNs, Transformers and decision trees. The interviewer went beyond basic definitions and asked about how these models work and when particular approaches would be appropriate. The decision-tree discussion went considerably deeper than I expected. I was asked several follow-up questions and had to explain the concepts using the terminology expected by the interviewer. The ML round was significantly more challenging for me than the DSA round. I was comfortable with some of the questions but struggled with a few of the deeper theoretical follow-ups. After completing the interview process, I did not receive an offer. The overall process was: OA → Technical/DSA Round → ML Round → Not Selected Interview date: December 1, 2023. My biggest takeaway was that an Applied Scientist interview requires preparation in two very different areas. DSA is important for clearing the initial technical stage, but the ML round requires genuine understanding of algorithms and deep-learning concepts rather than just knowing definitions.