40 interview questions on Machine learning - Fundamentals for developers (2025)

Hard

18 votes1216 runs0 comment

How to handle missing data in a given training data set ?

Data Scientist

machine learning

missing data

Martinat Apr 4, 2022

Medium

11 votes1182 runs0 comment

The validation set is used to provide frequent and unbiased evaluations of the model’s fit on the training set while tuning its hyperparameters/parameters: in other words,the model is found and then tested on the validation set before to be improved once again.

Data Scientist

hyperparameters

model

Martinat Mar 15, 2022

Hard

9 votes1100 runs0 comment

How to handle missing data in a given training data set ?

Data Scientist

machine learning

missing data

Martinat Apr 4, 2022

Easy

7 votes1126 runs0 comment

What is the missing word in the following sentence: Overfitting is the production of a model that corresponds too closely to the training data set and may therefore fail to fit other data sets and so fail to perform any reliable forecast.

Data Scientist

machine learning

overfitting

Martinat Mar 15, 2022

Hard

5 votes1076 runs

1 comment

What are the differences/similarities between a loss function, an error function, and a cost function?

Data Scientist

loss function

error function

cost function

Martinat Mar 15, 2022

Easy

5 votes742 runs0 comment

What is ensemble learning? It consists in …

Data Scientist

models

ensemble

Martinat Mar 15, 2022

Medium

4 votes1136 runs0 comment

Given a data set, which elements will help you to choose what model / machine learning method to use to create an efficient model?

Data Scientist

data

model

Martinat Mar 15, 2022

Easy

4 votes1065 runs0 comment

In order to get good results when implementing deep learning, one needs to prepare his data before using it. What action should be done during this preparation?

Data Scientist

data

preprocessing

Martinat Mar 15, 2022

Medium

4 votes1032 runs0 comment

How to assess a clustering model performances?

Data Scientist

metrics

clustering

Martinat Mar 15, 2022

Easy

4 votes1020 runs0 comment

Clustering is an unsupervised machine learning process which aims at automatically discovering natural grouping in the input data. When using this method, the developer is able to choose the number of groups he wants to create.

Data Scientist

machine learning

unsupervised learning

Martinat Mar 15, 2022

Medium

3 votes1060 runs0 comment

Which of the following statements is true when considering machine learning?

Data Scientist

machine learning

artificial intelligence

computer science

Martinat Mar 15, 2022

Hard

2 votes1061 runs0 comment

How to define reinforcement learning? Select the sentence that defines best this learning method.

Data Scientist

reinforcement learning

machine learning

Martinat Apr 4, 2022

Easy

2 votes1020 runs0 comment

How to define supervised learning? In supervised learning, the machine trains the model using a labeled training data set.

Data Scientist

supervised learning

machine learning

Martinat Mar 15, 2022

Medium

1 vote1067 runs1 comment

Which of the following actions can be used during the date pre-processing when performing regression or classification?

Data Scientist

data

preprocessing

Martinat Dec 13, 2023

Medium

1 vote1067 runs0 comment

What is the relation between Artificial intelligence (AI), machine learning and deep learning?

Data Scientist

machine learning

deep learning

artificial intelligence

Martinat Mar 15, 2022

Easy

1 vote1044 runs0 comment

What type(s) of learning can the following performance indicators be used?

Data Scientist

supervised learning

performance indicators

Martinat Mar 15, 2022

Easy

1 vote1019 runs0 comment

How to define unsupervised learning? In unsupervised learning, the machine trains the model using an unlabeled training data set.

Data Scientist

machine learning

unsupervised learning

Martinat Mar 15, 2022

Easy

1 vote351 runs0 comment

Cross validation is a technique used for evaluating ML or DL (deep learning) models. The method consists in training several models on subsets of the input data and evaluating them on the complementary subset of this same data. It aims at avoiding overfitting.

Data Scientist

overfitting

validation

Martinat Mar 15, 2022

Medium

1 vote311 runs0 comment

Which of the following statements are true when considering boosting?

Data Scientist

ensemble

models

Martinat Mar 15, 2022

Medium

1 vote307 runs0 comment

Select the true statements about PCA.

Data Scientist

machine learning

dimension reduction

unsupervised learning

Martinat Mar 15, 2022

40 interview questions on Machine learning - Fundamentalsfor developers (2025)

Test your technical skills before an interview of Machine learning - Fundamentals with 40 corrected questions. Beginner, Intermediate or Expert level.

How to handle missing data in a given training data set ?

The validation set is used to provide frequent and unbiased evaluations of the model’s fit on the training set while tuning its hyperparameters/parameters: in other words,the model is found and then tested on the validation set before to be improved once again.

How to handle missing data in a given training data set ?

What is the missing word in the following sentence: Overfitting is the production of a model that corresponds too closely to the training data set and may therefore fail to fit other data sets and so fail to perform any reliable forecast.

What are the differences/similarities between a loss function, an error function, and a cost function?

What is ensemble learning? It consists in …

Given a data set, which elements will help you to choose what model / machine learning method to use to create an efficient model?

In order to get good results when implementing deep learning, one needs to prepare his data before using it. What action should be done during this preparation?

How to assess a clustering model performances?

Clustering is an unsupervised machine learning process which aims at automatically discovering natural grouping in the input data. When using this method, the developer is able to choose the number of groups he wants to create.

Which of the following statements is true when considering machine learning?

How to define reinforcement learning? Select the sentence that defines best this learning method.

How to define supervised learning? In supervised learning, the machine trains the model using a labeled training data set.

Which of the following actions can be used during the date pre-processing when performing regression or classification?

What is the relation between Artificial intelligence (AI), machine learning and deep learning?

What type(s) of learning can the following performance indicators be used?

How to define unsupervised learning? In unsupervised learning, the machine trains the model using an unlabeled training data set.

Cross validation is a technique used for evaluating ML or DL (deep learning) models. The method consists in training several models on subsets of the input data and evaluating them on the complementary subset of this same data. It aims at avoiding overfitting.

Which of the following statements are true when considering boosting?

Select the true statements about PCA.