Talk about your research. Explain the attention mechanism and positional encodings of the transformer. Write down the formula of positional encodings. Explain graph laplacian and laplacian positional encodings for graph transformers.
Applied Scientist Interview Questions
1,159 applied scientist interview questions shared by candidates
Some basic ML questions and a lot LP questions.
Leetcode medium and machine learning 101.
Started with the projects listed in my resume. Then asked questions related to NLP(Transformer, RNN) and their architecture, model evaluation, working, embeddings etc.
DSA-An easy leetcode problem based on greedy algorithm.Usage of priority queue was there.sECOND QUESTION WAS OF MULTI SOURCE BFS.In the second round, they asked mainly on my projects and basic of ml like bias , variance trade off,random forest,some CNN architectures like ResNet, VGG etc.
1. Previous Work Experience 2. What is ResNet and how does it work 3. How to split node in Decision tree 4. What happens during runtime with dropout regularization 5. Full detailed mathematical formula for Wasserstein Distance!!!! - Very difficult to know/remember. Interviewer does not accept intuitive
Tell me about yourself, please
Propose ML methods for two real-world use cases presented by the interviewer, it also involved a live-coding challenge and additional ML-related theoretical questions
1) How to approximate pi using Monte Carlo? 2) What is the objective function of a neural network for classification and regression problems? 3) What are the different ML problems?
Describe your favorite project and then questions related to that project.
Viewing 261 - 270 interview questions