Question Bank
Machine Learning interview questions
Naming a model or a metric is only the start of an interview answer. Explain why it fits the task, and connect training, regularization, validation and evaluation to a concrete example. Start with your own explanation, then state its assumptions and limitations.
Showing 20 of 746 questions
Do you use AI assistants and LLMs in your work?
What types of candidate generation exist in transformers?
Is Python used as the main development stack?
Do ML engineers handle model deployment completely?
Understanding of task and infrastructure requirements
Explain bias-variance decomposition for decision trees and gradient boosting
How does IVF index work in FAISS?
How does implicit ALS (Alternating Least Squares) work?
How does Unbiased Learning work for ranking (possibly CatBoost)?
Have you worked with NDCG metric? How is it calculated?
Have you worked on model deployment? What tools did you use?
How do you approach selecting models for content generation?
How to process requests for describing generated images?
When comparing LLM with BERT, which model is better and why?
Does the search system work across the entire Avito database or only within specific groups?
What agent architecture is used for accessing internal company data?
How did you determine the question domain for AI assistant and product strategy?
How did you use filters with embedding-based search?
Prepare for your next interview with Vibe Interview.
Download Vibe Interview