Foursquare Data Scientist Interview Questions
& Process
Real candidates share what happened, how many rounds they had,
and how the experience turned out.
Based on 7 interview experiences · FREE TO READ
Candidate interview experiences
First-hand accounts from people who interviewed at Foursquare.
Foursquare
Data Scientist
Entry Difficult Positive experience No offer 1 round
Interview process
Technical screen
Interview formats
Coding Technical
There was a Hackerrank technical interview with an interviewer. It involved live coding on a general algorithm, and I could choose my coding language. They also asked a generic "tell me about yourself" before the live coding started.
Confirmed questions2 questions
- Can you write a function to find the maximum profit given a list of stock prices, where the purchase must occur before the sale?
- Could you share some information about yourself?
Foursquare
Data Scientist
Senior Difficult Positive experience Decline offer 5 rounds
Interview process
Recruiter call Technical screen Onsite Background check Offer
Interview formats
Coding Technical Case Presentation
The interview process was pleasant overall. Overall, I got the impression that foursquare is a company with interesting data sets, talented employees, and good WLB. It started with a hiring manager screen that had an open ended hacker rank coding question about location data and a few follow up questions. The onsite had four rounds: 1. Programming (implement a class to perform certain functions). 2. Statistics and ML (some standard questions on ML, as well as some finance/trivia probability questions). Rounds 3 and 4. Data science case study involving location data with a little bit of coding.
Confirmed questions7 questions
- Can you implement a class to perform data processing?
- What is the bias variance tradeoff?
- How would this approach scale to a large dataset or on hadoop?
Foursquare
Data Scientist
Mid Average Negative experience No offer 3 rounds
Interview process
Recruiter call Technical screen Onsite
Interview formats
Technical Coding Behavioral
The recruiter emailed, saying the team wanted to move fast, so we skipped the recruiter call. The email asked the usual HR stuff like location, hybrid, work authorization. Then there was a 60-min technical screen where we did a deep dive into a previous project, and I was asked a stats question about sampling from a population with replacement. They wanted me to dictate a proof and explain steps, which was weird and not really relevant. There was also a hackerrank open-ended coding question about parsing data files without pandas or numpy. The onsite had 1 sr data scientist, 2 software engineers, and 1 manager call with someone who wasn't the hiring manager. We did hackerrank coding questions again, parsing data without pandas, numpy, or sql. The interview process really doesn't align with the skills needed for the job; they said the non-pandas/numpy/sql questions were to see how candidates think, but then admitted pandas/numpy/sql are used daily. The data science problems they're working on seem more academic and research-focused, not so much on product strategy or business impact. The final straw was the sr manager saying they're looking for software engineers who can do some data science. The company has been around, offering a tech industry feel, but it's a red flag if many data scientists and managers have only worked there, hiring is disorganized, and data science projects don't clearly impact business strategy.
Confirmed questions8 questions
- Can you tell me about a previous project you worked on?
- Please explain the mathematical proof for sampling from a population with replacement.
- Solve a hackerrank coding problem regarding parsing data files, avoiding pandas and numpy.
Foursquare Data Scientist Interview Questions
Quoted word for word from Foursquare interview reports.
“Given training data with 1,000 points and a bootstrapped sample of 1,000 points (with duplicates), what is the expected number of unique points in the bootstrapped sample?”
Read reports →“What is the total count of unique devices from day 1 to day n?”
Read reports →“Solve a hackerrank coding problem regarding parsing data files, avoiding pandas and numpy.”
Read reports →“What is the bias variance tradeoff?”
Read reports →“Explain the reasons why XGBoost often performs better than other models.”
Read report →“Solve hackerrank coding problems involving data parsing without pandas, numpy, or SQL.”
Read report →“Can you write a function to find the maximum profit given a list of stock prices, where the purchase must occur before the sale?”
Read report →“Can you explain the bias variance trade-off concept in machine learning?”
Read report →“How would this approach scale to a large dataset or on hadoop?”
Read report →Formats, difficulty and experience
Across all 7 Foursquare interview reports.