If Amazon wants to send email to potential customers, how to model this problem? What features to use? Which metrics to evaluate? Which metrics (recall/precision) is more important? DL: - Describe the basic ANN. - What's the benefit of ReLU?
Applied Scientist Interview Questions
1,174 applied scientist interview questions shared by candidates
How would you learn an ML model where you did not have many labels or was sparse
They have asked various basic questions from machine learning and Deep learning , Recommendation Systems relevant. They also asked two design various models from Recommendation Systems. One round was online coding from basic data structures.
NDA, cannot disclose.
Explain ICA, and CCA. How do you get to CCA objective function from PCA.
What data structure would you need for BFS?
Lots of behavioral questions, e.g. how do you resolve a conflict?
Five people during "in-house": One algorithms/coding, breadth and depth ML (one each), one ML architecture (recommendation systems), one "leadership principles".
The technical screening involved some leetcode style problems and general problem solving + ML questions. The round table was a mix of behavioral questions, statistics, ML, and at least 1 more instance of leetcode problems. Question difficulties ranged from easy, like "how to address in the dataset", to average difficulty questions like "explain the differences between and another" or "can you describe how works.
Describe to me what up-sampling is in a convolutional neural network.
Viewing 91 - 100 interview questions