Machine Learning Engineer • Mid-Level
Machine Learning Engineer • Mid-Level
High QualityInterview Summary
Machine Learning Engineer coding phone screen at Reddit focused on building a classification pipeline using Python, pandas, and common machine learning models. The exercise involved predicting whether a user would click on a social media post based on their historical browsing activity and the category of the current post.
Interview Details
Round 1 — ML Coding: Predict User Clicks from Browsing Activity
The interviewer presented a supervised classification problem involving user engagement with social media posts. I was given a dataset containing information about users' historical browsing behavior and the category of the post they were currently viewing. My task was to build a machine learning model that could predict whether a user would click on that post.
Part 1 — Load and Explore the Dataset Using Pandas
The input data was provided in JSON format, and I needed to load it into a pandas DataFrame. Each record contained the following features:
hours_A: Number of hours the user spent reading posts in category A.hours_B: Number of hours spent reading posts in category B.hours_C: Number of hours spent reading posts in category C.category: Category of the current post, either A, B, or C.click: Binary target label indicating whether the user clicked on the post (1) or did not click (0).
The interviewer specified that the dataset had already been preprocessed. There were no missing values or class imbalance issues to address, so I could focus on the modeling and evaluation process.
Part 2 — Train and Compare Multiple Classification Models
After loading the dataset, I needed to build several classification
Unlock Full Details
Create a free account to unlock this interview experience report.