Machine learning is a technology that is breaking ground at new speeds every day. Technically, it should be improving faster and faster, given that ML is essentially supposed to be learning itself.

Modeling & Machine Learning Interview

Introduction

Overview

How Much Machine Learning?

Types of Machine Learning Interview Questions

Introduction to Machine Learning

The machine learning and modeling case study is the most common type of interview question that tests a combination of modeling intuition and business application.

Case Study Framework

Modeling Case Study

Data processing and analysis is the first step that we need to consider once we've clarified details and started down the path of building the model.

Missing Housing Data

Variable Error

Skewed Pricing

Bank Fraud Model

Data Pre-Processing

Feature selection and feature engineering is the second part of the data processing step. Once we've understood what our data looks like, we need to begin to theorize the kinds of features we would use to build the model.

Ride Requests Model

Multicollinearity in Regression

Keyword Bidding

Feature Selection

Model selection is usually the crux of any modeling case study problem. We want to be able to select a model or machine learning algorithm that will combine a bunch of factors to become the most optimal algorithm for the problem.

Booking Regression

Evaluate News

Credit Card Fraud Model

Model Selection

We have touched on the different machine learning algorithms throughout this lesson, but haven't yet dived deep into each one. The prior for this course is that you, as a candidate, have an idea of basic machine learning concepts, and the different modeling algorithms are one such example of them.

Machine Learning Algorithms Framework

Regression, Regularization, and Gradient Descent

Tree Algorithms

Coefficients of Logistic Regression

Encoding Categorical Features

Assumptions of Linear Regression

Machine Learning Algorithms

Recommendation and search engines are a subclass of the information filtering domain. They are used in almost every single modern web application or platform. Implementing recommendation and search engines usually requires a complex team of data scientists, data and software engineers, and machine learning engineers to collaborate and build at every company.

Comparing Search Engines

Job Recommendation

Newsfeed Model

Restaurant Recommender

Recommendation and Search Engines

Most machine learning model deployment requires some technical details and implementation to doing so. But we can abstract away from that in an interview when we’re focusing on the model roll out.

Cross Validation

Evaluation Metrics

Testing and Roll Out

Loan Model

Food Delivery Times

Spam Classifier

Delivery Estimate Model

85% vs 82%

Model Evaluation

Applied modeling is a type of case question asked about practical machine learning. The most common type of question framework is: Given an example scenario with a machine learning system or model, how would you analyze and fix the problem?

Rejection Reason

Active User Probability

One Million Rides

Parents Joining Teens

Applied Modeling

Machine learning system design focuses more heavily on the engineering aspects of model deployment. While slightly out of scope, we still wanted to cover some of the basics.

Type-ahead Search

Youtube Recommendations

Spotify Discover Weekly

Dynamic Demand Pricing

Machine Learning System Design

Regression models are used to predict the value of a dependent variable from one or more independent variables.

Introduction to GLMs

Linear Regression

Lasso vs Ridge

Using R Squared

Logistic Regression

Perfectly Separable

Softmax vs Logistic

Time Series Model

Why Do We Need Time Series Models?

AR and MA

Significance Time Series

Generalized Linear Models and Regression

Understanding of how to build, deploy, and test machine learning models and tackle business problems while applying modeling concepts. 

Learn how to apply machine learning to influence business decisions

<p>How would you interpret coefficients of logistic regression for categorical and boolean variables?</p>


<p>Why is one-hot encoding recommended for categorical variables in logistic regression models?</p>

Why is one-hot encoding recommended for categorical variables in logistic regression models?

To avoid the model assuming an inherent rank or structure to the different categories

To convert categorical variables to continuous variables

To increase the magnitude of the coefficients

To ensure all variables have the same scale

To convert categorical variables to boolean variables

Modeling & Machine Learning Interview

Coefficients of Logistic Regression
Go to question details page

31%

Modeling & Machine Learning Interview

Coefficients of Logistic RegressionGo to question details page

31%

Coefficients of Logistic Regression
Go to question details page