Majority questions were around tabular machine learning and data understanding.

Use of Dropout while training a model and once trained, how to use it during test process. Complete ML workflow process: The techniques used for Preprocessing a dataset Which model to use (based on Classification, Regression etc. problem) Training, Validation and Testing Deployment Search for questions that can be asked in similar scenario Classification/Regression algorithms in detail: Which algo to use in a specific use case and why? Transformer architecture Bias, Variance, Overfitting, Underfitting: Ways to overcome Under and Over fitting Scenario: If a Model shows a graph with a lot of variations (up and down), it refers more towards underfitting as model is unable to understand the comple and data and is making assumptions. Part 2 Good Python concepts: Typecasting args and kwargs yield FastApi Database: How to handle large datasets. User provides their data and how you will create a profile from this. Profiling means # of columns/rows, empty values, other stats about the data