Machine Learning / AI

What criteria would you use to evaluate the effectiveness of a recommendation system that takes tasks from a database or generates them via LLM?

You go too deep into the technology when asked to talk about simple things. How would you comment on this?

Suggest an option for developing a task recommendation system for a student on an educational platform.

How to ensure the relevance of information in an RAG system during constant document updates?

What is the difference between overfitting and underfitting?

What happens to list tails after the main loop and how to process them?

What should you do if you don't have a validation set? How to validate a model?

Tell about yourself, your experience, what tasks you have had to deal with.

What are L1 and L2 regularization? How do they help prevent overfitting?

What is data normalization and why is it important in model training?

How would you explain the recommendation system to a physical education teacher (non-technical user)?

Can data leakage occur when normalizing data before splitting into train and test?

How would you understand why students in a group with a recommendation system stopped performing the recommended tasks?

What is data leakage and where does it come from?

When is it better to use RAG, and when to use model fine-tuning?