Me postulé a través de un reclutador. El proceso tomó 2 semanas. Acudí a una entrevista en Amazon (Seattle, WA) en jun 2025
Entrevista
I had a mixed feeling after the interview itself that they wont pass me to next. I only had a phone screening, where HR said to only prepare for Leetcode/coding, but the interviewer was all over the projects I did, ML/DL, probability related easy coding question, but nothing related to data structure or algos, so doing leetcode 150 was also like a waste. The interveiwer was very specific about the partoicular word he wish to hear and nothing like other factors which can effect the final outcome of model. I asked him to share the screen and show my paper to explain it better, but he refused to share screen saying that in this part we usually dont allow candidated to share screen.
Preguntas de entrevista [1]
Pregunta 1
Project discussion, why not chosed some specific baselines which were naive, and philosophical questions like why the world is moving away from CNN or VAE or those architectures and coming to transformers like ViT, Catastrophic fogetting happens only due to LR in NN, which is a partial truth, Given PDF and RV, draw some specific number of points from it with a given list of probabilities, ROC-AUC, Evaluation metrics, Clustering algos, classical genAI algos, Dont quote anything mathematically but theoretically
Me postulé a través de un reclutador. Acudí a una entrevista en Amazon (Seattle, WA) en jun 2026
Entrevista
This interview was for the Applied Scientist Position at Amazon. It was a Science breadth/Depth as well as 1 easy LC problem. After that did a full loop interview. 1 Science ML Breadth, 1 Science Depth, 1 Sys Design (Team-specific), 1 coding round and 1 Bar Raiser
Preguntas de entrevista [1]
Pregunta 1
Bias-Variance tradeoffs. Bagging Vs boosting. Modeling details in my project. Basic Statistical Questions. DSA: String compression problem.
1 HR round
4 technical interviews (coding+depth+breadth) of Machine Learning.
1 round to go into depth of my own projects.
1 round on general data science questions + system design (model a pipeline end-to-end for translating one set of multimodal objects to another language)
Acudí a una entrevista en Amazon (Santa Monica, CA)
Entrevista
Two easy Leetcode problems, some ML coding problems and a lot of Leadership Principles talk. Overall average experience. A lot of middle managers not much room to innovate or take risks. Like working for the DMV