Machine Learning / AI

Pre-layer norm vs post-layer norm: what is the difference, where are they used and why?

What will be the result of two SQL queries working with the departments and employees tables?

What are temporal knowledge graphs?

What string recognition models are there (CRNN, TrOCR)?

How does DiT (Diffusion Transformer) differ from U-Net in diffusion?

Senior

HybridОстровок

How does tool calling occur in the agent?

Middle

СБЕР

What roles were in the team?

Which is preferable from a computational graph perspective — dynamic or static?

Middle

Иннотех

What is multi-query attention and where is it used?

What are MAPE and SMAPE metrics? Where might they be ineffective?

How to solve the problem of summing revenue on a large volume of data?

What is global average pooling and why does it replace a fully-connected layer?

How familiar are you with the outstaff format?

What is satellite imagery and what are the sources (Sentinel, Landsat, Maxar)?

Senior

HeadHunterГазпромбанкЛеруаМерлен

Что будет, если вызвать max вместо idxmax для поиска максимального значения?

Middle

СБЕР

From what date are you ready to start cooperation in case of a positive decision?

What is a branching criterion in trees?

Middle — Senior

Яндекс

What are stride and padding? How do they affect the output size of a convolution?

Junior — Senior

Домклик

How to parse documents for RAG?

How to implement a query with the maximum exchange rate date for each transaction considering weekends and holidays?

Middle

OZON