Find jobs for junior data engineer

Junior data engineer Interview Questions

705

Junior Data Engineer interview questions shared by candidates

Top Interview Questions

Sort: Relevance|Popular|Date
Allstate
Junior Data Scientist was asked...January 5, 2017

If a jar has X red balls and Y blue balls, what is the minimum number of draws that is necessary to ensure that you have one ball of each color?

3 Answers

max(X, Y)+1

It is max, the first reply was correct. I you have 5 blue balls and 2 red balls : - If it was min : you take 3 balls, it is still possible to have picked 3 blue balls. - with max : you take 6 balls, the worst case is that you have picked 5 blue and 1 red balls. Less

min(X,Y) + 1 is correct, not max()

Automatic

If the coefficient of an independent variable in linear regression has an opposite sign than it should be, what might be the reason

1 Answers

multicollinearity

Actori

Introductory questions. Describing yourself. Your skills. Your past experiences. Presentation of a Case Study that was provided a few days before the interview.

1 Answers

I answered honestly. I sounded genuine and friendly. Smiling all the time. Presented the Case Study and answered the questions that followed. Less

Tell us about your programming learning experience?

1 Answers

Talked about my learning experience

Are you a challenging person?

1 Answers

I said yes

Schwarz Dienstleistung

Wann benutzt man Decision Trees, wann Neuronale Netze? Wenn ein neuronales Netz 90% Accuracy auf dem Training Set hat, aber 50% auf dem Test Set, was haben wir dann? -> Overfitting ..Basics zu Machine Learning

1 Answers

Wann benutzt man Decision trees: Um herauszufinden, ob das Problem ML überhaupt benötigt und /oder um eine Benchmark zu bekommen, die das NN schlagen sollte. Unterliegt das NN, weiß ich, dass ich den Data Scientist rausschmeißen kann, da er Fake ist. Wenn ich ein modernes Framework (z. B. Tensoflow) verwende, sind NN den DT in jeder hinsicht überlegen und bieten keinen Vorteil, außer vlt. kürzere Trainingszeit unter bestimmten Bedingungen. Das ist aber vernachlässigbar. Accuracy ist im Training höher als Test: Es kann natürlich Overfitting (das NN lernt die Trainingsdaten auswendig) sein. Genauso kann aber ein Auto, dass nicht anspringt entweder einen Motorschaden oder einen leeren Tank haben. Man müsste erstmal die Lernkurve ansehen (z. B. TensorBoard) und dann einen Blick auf die Daten werfen: Sind die Daten vlt. noch sortiert? Sind die Daten in Test und Training gleich verteilt? -> Ich mache [train, test].shuffle und splitte danach, was für Werte habe ich nun Habe ich Features die zu stark gewertet werden? -> sind IDs drinnen oder zu genaue Timstamps, habe ich alle Werte normalisiert Welche Architektur hat mein NN? Was passiert wenn ich verschiedene Parameter durchprobiere (Hyperparameter Tuning)? Less

Enhance IT

Given a list of words, sort these words from lowest character length to highest.

1 Answers

Just make sure to know the common sorting methods in your programming language of choice. Less

Allstate

If you have a bag of jelly beans with a large number of two different colors, whats the minimum you need to pick out to guarantee that you have at least 2 of the same color? Then the question in the general case.

1 Answers

3, and for the general case picking m with the same color from n classes you'd need: m * n + 1. Use "worst case scenario" analysis. Less

Allstate

Taje a jar with stones of three colors, how many draws do you need to get two stones of the same color? Generalise to n colors, k stones.

1 Answers

Answer is 4 draws and the formula is n(k-1)+1 = 3(2-1)+1 = 4

A1 Hrvatska

What was your favourite subject at college?

1 Answers

Machine learning

Viewing 1 - 10 of 705 Interview Questions