What is the cold start problem (user/item) and how is it solved?
Machine Learning / AI
Is there a team lead in the team?
If the needed chunk is at the 10th position out of 10, and only 3 can fit into the context — how to evaluate and how to raise it to the top 3?
Why won't a linear model without feature transformation solve the problem of separating points inside and outside a circle?
What to do if the request does not relate to any scenario at all (for example, 'hello' or spam)?
Why do you want to leave [organization] and what does professional development mean to you?
Why can comparing float numbers be incorrect and how to deal with it?
Why is F1-score used instead of average precision and recall?
Tell about your project with two models (GigaChat and a local model trained with LoRA)
What threats needed to be detected in a video surveillance system?
What is an instrumental variable?
What is the time complexity of adding and removing an element in a Python dictionary?
How do you account for duplicated points when checking for symmetry?
What is hybrid search and what is its main idea?
How to use cross-validation to evaluate a model?
What will be the result of generating synthetic data for a rare class? What will happen to the distribution?
What is A/B testing for ML models and what business metrics are important?
In what form does LLM classify the query — what exactly does it return for scenario selection?
What is Kafka?