DREAM OF BECOMING A DATA SCIENTIST?
Check Your Skills with Fayrix Instant Test!

When you apply for a position of a Data Scientist, you take risk of landing to an interview, where the HR manager challenges you with a ton of abstract and barely relevant questions, which can hardly assess you as a good candidate. No one knows the reasons for that, perhaps this is caused by the lack of real Data Science competences or efforts to earn respect.

We conduct a lot of interviews in Data Science. We @Fayrix look for really cool and gifted employees. In order to separate talented candidates from medium ones, we have to look for non-standard tasks not to be boring and leave pleasant impression from the interview.

We have introduced this interview task a year ago for the first time and still remember those faces, silently asking "pictures are, of course, fun, but how to solve the task is not clear".

Now you can try and test yourself for being an outstanding Data Scientist.

So, here is the deal. Look at the picture and choose the most appropriate mapping order for the following key terms: XGBoost, linear regression, Kaggle, Confusion Matrix.

Answer options:

1. A, B, C, D
2. C, B, A, D
3. C, B, D, A
4. A, B, D, C

Let's start with the easiest - the picture with a pregnant man (False Positive) from the presented options refers to the Confusion Matrix (Error Matrix). Comedians can see kaggle in it - "do not believe anything, in the competition everything will turn out to be different from what I was taught".

Now let's try to group the proposed options:

• XGBoost is a method.
• Linear regression is a family of methods.
• Kaggle is a competition platform.
• Confusion Matrix is a term.

XGBoost and Linear regression are popular methods, and if you think about the relationship between them, most will say that XGBoost gives a better result, i.e. cooler. If you look for a similar relation among the remaining three pictures, you will easilty find the correct answer - a car and a helicopter.

Using the exclusion method, we conclude that kaggle is depicted by the picture with a tower of cubes symbolizeing the popular technique - stacking.

So the correct answer is 2. C, B, A, D.

For skeptics who don't find this test accurate enough, I should note that the correct solution is mainly based on the candidate's overall erudition, as the picture of the car-helicopter pair is borrowed from a popular article about XGBoost, and the tower of cubes is a popular meme in a very well-known community of Data Scientists.

And of course, the task narration has a very important accent "Choose the most appropriate order". For me personally, tasks that can have several correct solutions are much more fascinating than those with a single defined answer. Such questions imply an open answer and invoke a dialogue demonstrating how the candidate thinks, their path of logics.

What other interpretations are possible, you would ask? The tower of cubes may be associated with XGBoost, it is a composition of trees, then the helicopter can depict kaggle, as something very modern, but now you face to associate regression with some of the pictures.