How do you handle the situation when a product is sold by different sellers?
Machine Learning / AI
So, you haven't had an end-to-end development cycle from data collection to deployment at any of your workplaces?
What methods of experiment clustering (cluster randomization) are there?
Have you deployed your models to production yourself?
Have you worked with Docker, MLflow, Airflow?
What are the limitations of AutoML on large data?
How does a convolutional layer work?
Tell me about the transformer architecture: encoder-decoder, what happens inside the decoder?
Name the classic methods of obtaining word embeddings (Word2Vec, FastText, TF-IDF, N-grams, etc.).
Tell us about Batch Normalization: what it is and why it is needed?
What is the difference between OLTP and OLAP systems?
Why does L1 regularization tend to zero weights, while L2 does not? How is this mathematically derived?
Do they select a subset of features in individual tree nodes or for the entire tree?
Tell about Random Forest: how the training works, how trees are formed, and how predictions are made for classification tasks.
What are the manifestations of weak models?
How to organize work starting from the formulation of tasks?
What is a chat template and why is it needed for instructional models?
How to organize data auto-collection and training using Airflow?
What is survival analysis and what data is needed?
What is bagging?