Featurespace Mid-Level Data Scientist | January 2022

Company: Featurespace

Location: Cambridge, United Kingdom

Role: Data Scientist

Level: Mid-Level

Outcome: Rejected

When did you have this interview?: Last month

Time:30 Mins.

It was though.

1- Tell me about yourself – up to 5mins

then technical questions started.

What kind of technical questions did you get asked?

2- Let’s talk about your recent project in which you used K-means clustering, could you explain how to evaluate your model?

3- How do you evaluate your model performance?

4- Could you explain to me one of your favourites / applied algorithms? And What are the main parameters of the method? (Decision Tree type method)

5- What are the differences between Random Forest and XGBoost?

6- What if training data performs perfectly, and test data fails, what is this situation called, explain it?

7- Once the overfitting happens which parameters would you tweak that will fix the problem on Decision Tree-based methods?

8- 5 Minutes CODING QUESTIONS (Explain at least pseudo-code style) - HIGH PACE- a. Given an array; we know it is incrementally increasing A= [1,2,5,9,13,17,21,25, 62, ….,1231] and we are looking specific number like x = 24. How do you write your method? (What would be your )

b. Do you know Big O Notation?

c. Imagine the previous array is a thousand times bigger; A.Shape = 1 x 1.000.000 How do you find your target this time x = 24293?

d. Ok, multiple your array 10 times so how can you handle the time difference for a 10.000.000 size array?

9- Last 5 MINUTES SEPERATED just for YOUR QUESTIONS. They didn’t tell anything until I ask a question