Amazon

I applied for the Applied Scientist Intern opportunity at Amazon through the off-campus process. The interview process had two mandatory technical rounds, and both had to be cleared to move forward. The first round was essentially an SDE-style technical interview focused on DSA. The interviewer started with a short introduction and then moved directly into problem solving. I was given a coding problem and was asked to explain my approach before writing the implementation. The discussion focused heavily on how I broke the problem down, the brute-force approach, and how the solution could be optimized. After reaching the main approach, I was asked about time and space complexity along with a few edge cases. There were also follow-up changes to the problem, so I had to explain how my solution would behave if the constraints were modified. The round felt quite similar to a standard Software Development Engineer interview rather than an ML-specific discussion. The main emphasis was on DSA, logical reasoning, optimization, and coding. I was able to work through the problems and complete the technical discussion. The second round was completely different and focused on Machine Learning. This round was also mandatory, so clearing the first SDE round alone wasn't enough. The interviewer asked questions around fundamental ML concepts and how different algorithms work. The discussion covered topics such as model training, overfitting and underfitting, evaluation metrics, and the reasoning behind selecting one approach over another. I was also asked to explain concepts rather than simply provide definitions. Some questions were scenario-based, where I had to think about what I would do if a model performed poorly or if the characteristics of the dataset changed. The interviewer was interested in the reasoning behind my choices and how I would approach an ML problem from a practical perspective. There was also discussion around my ML-related knowledge and projects. I had to explain the approach I had taken, the algorithms involved, and some of the decisions made during implementation. The two rounds required quite different preparation. The first one was primarily about DSA and software-engineering problem solving, while the second shifted completely toward ML fundamentals and analytical thinking. Since both rounds were mandatory, preparation for only one side of the role wouldn't have been sufficient. After completing both rounds, I waited for the final update. Eventually, I received the information that I was not selected for the Applied Scientist Intern position. The main takeaway from the process was that an Applied Scientist interview can still involve a strong software-engineering component. The first round tested coding and DSA at an SDE level, while the second evaluated whether I had the ML foundation expected for the role. Preparing for both sides was therefore essential.