Банк вопросов из реальных собеседований

Why even with FlashAttention, learning on hundreds of thousands of tokens might not fit into GPU and what classes of solutions reduce memory and computing?

What do you need to check in character building and time series partitioning to find a leak of future data?

One of the heads of the multi-metal model works poorly. How to check signs, markup, temporary processing and collect useful complex examples?

Есть много неразмеченных driving/log sequences и мало labels для редких событий. Какие self-supervised подходы можно использовать до supervised fine-tuning?

How does scalable scalar attention work, what is its computational complexity, and what changes GQA, MQA, and the limited attention window?

The company builds text profiles of users by their history using a generative model. How to safely use such profiles in search and ranking?

Design a system that searches the internet for potential brand violations and filters out a ton of irrelevant results for 1000+ customers.

How to formulate a target variable for forecasting the future movement of the average price on trades and the exchange glass?

Whisper performs poorly on a target language such as Vietnamese or Uzbek. How would you adapt the system and collect data for this appointment-extraction task?

How to explain the complexity of the algorithm and the formal definition of O-large in terms of constants and input size?

How can H-index be computed in O(n), why may citations be capped at the number of papers, and what changes if the input array may be modified?

At the interview they asked: how many zeros are at the end of the number 100!, and how to accurately calculate this without calculating the factorial itself?

At the interview they asked: how many zeros are at the end of the number 100!, and how to accurately calculate this without calculating the factorial itself?

How do you reason about algorithmic complexity, when is a greedy choice valid, and how can an unknown biased coin produce a fair random bit?

When to use async, threading and multiprocessing in Python, and how does the GIL influence these choices?

Compare bagging with boosting, explain what a p-value means, and show how Bayes’ rule updates a hypothesis after an observation.

You need to verbally design a simple in-memory vector search: add, search top-K, cosine similarity, stats. What do you pay attention to?

What to say about linear programming, the simplex method and greedy algorithms if asked at a technical ML interview?

What basic runtime questions for Python often come after an algorithmic problem?

Describe the reservoir sampling algorithm for one element from a stream and explain why each element seen is selected with equal probability.

Compare REST and gRPC at a high level. Then explain what a database index does and what simple data structures can back an index.

The moderation model requires classes and data. How to collect labels, handle imbalances, and not mix different policies into one noisy dataset?

How are matrix equation, least squares, gradient descent and L1/L2 regularization related?

How does the distribution of selected numbers change if for each query we select the number with the maximum XOR?