The questions ranged from basic topics like bias vs. variance trade-offs to advanced topics like LLMs and self-supervised learning. Usually, they will give you a hypothetical scenario and ask you to solve the problem using ML. Based on your answer, you are asked subsequent questions.