Coding exercises
Sr Data Scientist Interview Questions
3,371 sr data scientist interview questions shared by candidates
experiments and key metric related q
- Diagram a solution for a multi-modal document prediction task. - Explain overfitting in the context of various algorithms (SVM, random forest, neural network) and techniques to mitigate it (very technical on the details) - Explain how you would deal with class imbalance or other sources of bias in a training data set (again, very technical on the details) - Explain how you would use a Docker container to make a deployable machine learning service using Kubernetes (wanted many devops details). - Have you used a pipeline automation tool like Luigi? Have you used monitoring tools like PagerDuty?
schema based SQL questions purely depends upon your understanding.
Explain a project you worked on. Then asked follow up questions on data and what techniques I used.
1. How is word transformers different from Sentence transformers. 2. Scenario based questions. 3. When to use F1 Score, Precision, Recall Score 4. Clustering before train test split or after train test split. 5. I was surprised on this question. Do SMOTE increase your test data sample size as well 6. Scaling - MinMax, Standard and Robust Scaler 7. When to use class weights and oversampling/Undersampling ( No one can get it right at the first time). I wont recommend for someone who has good experience in Data Science. In my decade of taking/giving data Science interviews, this has to be rated as my worst experience.
I was asked to explain some different ML models and compare them.
I was asked to talk about a project I am proud of.
Let's say you have worked on binary classifier random forest model, and both the classes recieved equal number of votes. How will you proceed further, what is your logic behind it?
What is a window function
Viewing 931 - 940 interview questions