Machine Learning / AI

Why were these particular features selected?

How did you choose the threshold of 0.6?

Why was CatBoost used?

Was the conversation compared in its entirety or only parts of it?

Have you made solutions for clients of your employer — did they take and support them themselves?

How did you start working on the project to evaluate dialogue quality? What data did you use, and what approach did you choose?

Dialogues were of different lengths — how did you handle this? Did all conversations fit into the context window?

How did you collect data and what approach did you choose for the churn task?

What were the signs of the employee?

How was the evaluation conducted? What metric determined whether it was a good or bad dialogue?

What projects and tasks does the team have?

Was there a baseline that was compared?

Why was this approach chosen (SentenceTransformer + cosine similarity)?

Tell me about a classic machine learning project — predicting employee churn. What was the business problem?

Was there data preparation, special character cleaning, or was it just what the operator said?

Understanding the smart assistant's task — a two-step classification

Tell me about your work experience

Which project do you find most interesting and want to tell about?