Facebook Data Scientist Interview Questions | Glassdoor

Facebook Data Scientist Interview Questions

Updated May 23, 2018
144 Interview Reviews

Experience

Experience
52%
24%
24%

Getting an Interview

Getting an Interview
37%
25%
24%
8
2
2
2

Difficulty

3.2
Average

Difficulty

Hard
Average
Easy

144 Candidate Interview ReviewsBack to all Interviews

Filter

Sort: PopularDateDifficulty

Filter

Sort: PopularDateDifficulty

Helpful (347)  

Data Scientist Interview

Anonymous Interview Candidate
No Offer
Neutral Experience
Difficult Interview

Interview

One HR interview, one takehome data challenge, one shared screen with SQL and one onsite with several 1:1 interviews. They check your coding skills and product sense via the takehome, your sql skills via the shared screen interview and machine learning theory as well as product sense during the onsite.
They let you choose the language for the takehome and onsite there is no coding on the board. So you just need to know one language (whichever you want, although I think they prefer R or Python) + SQL. No C++/Java/etc stuff and no CS algo questions.

Interview Questions

  • Data challenge was very similar to the ads analysis challenge on the book the collection of data science takehome challenge, so that was easy (if you have done your homework).

    SQL was: you have a table where you have date, user_id, song_id and count. It shows at the end of each day how many times in her history a user has listened to a given song. So count is cumulative sum.
    You have to update this on a daily basis based on a second table that records in real time when a user listens to a given song. Basically, at the end of each day, you go to this second table and pull a count of each user/song combination and then add this count to the first table that has the lifetime count.
    If it is the first time a user has listened to a given song, you won't have this pair in the lifetime table, so you have to create the pair there and then add the count of the last day.

    Onsite: lots of ads related and machine learning questions. How to build an ad model, how to test it, describe a model. I didn't do well in some of these.   23 Answers

Other Interview Reviews for Facebook

  1. Helpful (225)  

    Data Scientist Interview

    Anonymous Interview Candidate in Menlo Park, CA
    Declined Offer
    Negative Experience
    Difficult Interview

    Application

    I applied online. The process took 6 weeks. I interviewed at Facebook (Menlo Park, CA) in February 2017.

    Interview

    I applied online in late December and then spoke to a recruiter for about 15 minutes in early January. I was scheduled for an in-person interview in mid January where I interviewed with one data scientist in a 1:1 interview for an hour for the first round. Another recruiter then called me for a 30 minute prep for the 2nd round. The final, 2nd round interview was about 5 weeks later in late February, with 6 data scientists over 4 hours in the afternoon. I got an offer but it was for a lot less than I was expecting, and we couldn't bridge the gap enough for it to be worth it.

    Everyone seemed relatively nice, although I could tell that a lot of the questions are really designed to trip you up, like they want you to miss some detail or edge case. My advice would be to pay attention to every little bit of minutiae regarding the question, make sure you're staying on task, write on the whiteboard, and explain your thoughts. Industry word is that data science at Facebook is not what it once was and is more of a product data analyst role now, so make sure you're really into Facebook products because that's what you'll be analyzing.

    Also, I didn't get a single question about dice, cards, or any other brain teaser type questions. All these mentions of NDAs are missing the point of Glassdoor, people can be a little more verbose than "various questions". You can't trademark an interview question or claim it's a trade secret.

    Interview Questions

    • How would you measure the health of Mentions, Facebook's app for celebrities? How can FB determine if it's worth it to keep using it?

      If a celebrity starts to use Mentions and begins interacting with their fans more, what part of the increase can be attributed to a celebrity using Mentions, and what part is just a celebrity wanting to get more involved in fan engagement?   6 Answers
    • There is a table that tracks every time a user turns a feature on or off, with columns user_id, action ("on" or "off), date, and time.

      How many users turned the feature on today?
      How many users have ever turned the feature on?
      In a table that tracks the status of every user every day, how would you add today's data to it?   9 Answers
    • If 70% of Facebook users on iOS use Instagram, but only 35% of Facebook users on Android use Instagram, how would you investigate the discrepancy?   7 Answers
    • How do you measure newsfeed health?   3 Answers
    • If a PM says that they want to double the number of ads in Newsfeed, how would you figure out if this is a good idea or not?   5 Answers
    • We have two options for serving ads within Newsfeed:
      1 - out of every 25 stories, one will be an ad
      2 - every story has a 4% chance of being an ad

      For each option, what is the expected number of ads shown in 100 news stories?
      If we go with option 2, what is the chance a user will be shown only a single ad in 100 stories? What about no ads at all?   14 Answers
    • How do you map nicknames (Pete, Andy, Nick, Rob, etc) to real names?   5 Answers
    • Facebook sees that likes are up 10% year over year, why could this be?   7 Answers
    • How many high schools that people have listed on their profiles are real? How do we find out, and deploy at scale, a way of finding invalid schools?   4 Answers

  2. Helpful (4)  

    Data Scientist Interview

    Anonymous Interview Candidate
    No Offer
    Neutral Experience
    Average Interview

    Application

    I applied through a recruiter. I interviewed at Facebook in May 2018.

    Interview

    Got contacted by the recruiter on Linkedin. I was not actively looking for a new job but was just curious about their interview process so I responded. The recruiter did a quick phone screen asking about my background and gave some info about the role. She then scheduled a technical interview with a data scientist.

    The technical interview questions were pretty similar to what are mentioned on Glassdoor. First part is some sql questions where he typed the sample data and asked me to write queries to get the answers. The second part was a series of product related questions about facebook videos and groups.

    I agree with others that the interview process feels a bit impersonal. The interviewer looks like he was just performaing a routine and demostrated no interests in my background. For example, he asked me to describe a recent project I worked on and gave zero reaction after my response. I feel like he also had the 'correct answers' in mind for the product questions which should be more open-ended. It doesn't seem like they are looking for the type of data scientists that you'd expect (someone with experience in machine learning, modeling, data engineer etc), since there are other teams for that. The analytics data scientists just support the PMs by doing a lot of sql. It should really be called a data analyst or product analyst role.

    Interview Questions

    • Given 2 tables, one with the phone numbers that facebook sends the confirmation message to and another one with the phone numbers that confirmed the verification, write a sql query to calculate the confirmation percentage.   1 Answer
    • How would you measure the health of facebook group? What metrics and counter metrics would you use?   Answer Question
    • How would you measure the performance of a new feature that makes it easier for users to upload videos.   1 Answer
  3. Helpful (1)  

    Data Scientist Interview

    Anonymous Interview Candidate
    No Offer
    Positive Experience
    Average Interview

    Interview

    Recruiter contacted me out of the blue on LinkedIn, set up a quick chat on the phone. He proceeded to ask basic stats questions which I was totally unprepared for unfortunately.

    Interview Questions

    • Facebook has personal information such as gender and height. If I were to make a claim that men were taller than women, how would you go about proving/disproving this claim?   2 Answers
    • What is a Z-test? When would you use a Z test over a T test?   Answer Question

  4. Helpful (7)  

    Data Scientist Interview

    Anonymous Interview Candidate in New York, NY
    No Offer
    Positive Experience
    Difficult Interview

    Application

    I applied through a recruiter. The process took 1 day. I interviewed at Facebook (New York, NY) in April 2018.

    Interview

    The interview process was well structured. It followed exactly as recruiter explained in the email (and as mentioned on glassdoor).
    There was one SQL question and one product sense question. I was able to complete the SQL question (could have done it faster), but didn't provide a clear, convincing answer for the product question

    Interview Questions

    • SQL question:
      given a table of interaction between users (user_a | user_b | day), find number of users who had more than 5 interactions yesterday (assume there is only one unique interaction between a pair of users per day).

      Product Question:
      A user satisfaction survey was conducted for two groups of facebook users (each with 50 k sample size).
      Group1: who had enabled certain login security features
      Group 2: who had not enabled these security features.
      It was found that user satisfaction with group1 was 30% lower than with group 2. Why do you think so? Comment on how the survey was conducted?   5 Answers

  5. Helpful (1)  

    Data Scientist Interview

    Anonymous Interview Candidate
    No Offer
    Negative Experience
    Difficult Interview

    Interview

    The interview process was smooth but impersonal. The vibe at Facebook is unfriendly -- this is surprising for a company priding itself on connecting people and friends.

    The onsite has a few business case components coupled with a SQL and probability test.

    Interview Questions


  6. Helpful (1)  

    Data Scientist Interview

    Anonymous Interview Candidate
    No Offer
    Positive Experience
    Easy Interview

    Application

    I applied through college or university. I interviewed at Facebook.

    Interview

    Initial interview was relatively simple. Just questions about what aspects of available data you would use to answer certain business questions or anomalies. One relatively complex SQL question. Should've brushed up more on my SQL! What you come to find out is that the "Data Scientist" role with Facebook is really just more of a Data Analyst type role.

    Interview Questions

    • Say you notice a surge in average activity among users but the number of users is down, how would you explain this?   2 Answers

  7. Helpful (6)  

    Data Scientist Interview

    Anonymous Interview Candidate
    No Offer
    Neutral Experience
    Difficult Interview

    Application

    I applied through an employee referral. The process took 4 weeks. I interviewed at Facebook.

    Interview

    One call with the recruiter and one phone interview with a Data Scientist and moved on to on-site interview with 5 rounds of interviews. In the phone interview, it's mostly brainstorming metrics and an SQL query question. The 5 rounds of on-site interviews includes 2 SQL, 2 A/B testing and 1 Math interviews.

    Interview Questions

    • If there's a new feature that attracts more pop stars to use it, how do you evaluate whether this feature is successful or not?   Answer Question
  8. Helpful (2)  

    Data Scientist Interview

    Anonymous Interview Candidate in Boston, MA
    No Offer
    Negative Experience
    Average Interview

    Application

    I applied through a recruiter. The process took 2 weeks. I interviewed at Facebook (Boston, MA) in March 2018.

    Interview

    HR reached out to me on linkedin for data scientist infrastructure position. Scheduled an interview after 2 weeks. This was a technical round with sql, python and statistics questions. The interview lasted for an hour with live coding. The interviewer was very nice and i answered all of the questions. I was not sure about logic for one of the questions on python. Rest were right. I did not make it to the next round. I am sure what went against me but assume python was the issue.

    Interview Questions

    • A table has a column with full names(e.g. 'Tom Hardin'). write a python code to get distinct first names(i know its super easy in sql)   3 Answers

  9. Helpful (1)  

    Data Scientist Interview

    Anonymous Interview Candidate
    No Offer
    Neutral Experience
    Difficult Interview

    Application

    I applied in-person. The process took a week. I interviewed at Facebook in March 2018.

    Interview

    Was contacted by a Facebook recruiter and then the resume exchanged hands when finally got a call from Infrastructure team. Recruiter herself asked screening questions from sql, statistics and programming, trivial ones.

    It was followed by a phone interview. It was from 3 areas in sql, programming and machine learning. Recruiter was clearly interviewing to test the limits. There were 3 questions on sql, 1 on programming and 2 questions on machine learning concept. The focus on machine learning was quite less and primary focus seemed to be on sql and programming.

    Sql questions were on group by and joins. The first question needed a self join which i didn't get at first, but the interviewer was helpful. The other 2 questions were straight forward group by and join questions .

    programming question was a bit unexpected for doing in 20 mins. Was given a log file of format
    id, accepted, id1,id2,id3
    id, rejected, id5
    id3,request, id6

    and was asked to create a network graph from it. I guess the interviewer wanted the result in simply a dictionary , but started with what will be the best data structure to store it, to which i answered bi directional graph and then asked me to code the graph, which got me confused b/w the linked list approach and the normal dictionary approach and ended up taking more time. In midway he said that he understood my approach and moved to machine learning.

    In machine learning the questions were between bias variance trade off and PCA . In PCA he went into detail as in what method is used to decompose matrix, what is the significance of first principal component and in the end what is the relationship between the first and second principal component. I got stuck at the relationship between first and second principal component.

    I got rejected. My best guess is because of the programming question. Recruiter had suggested that phone interview will be around basic data processing question, so wasn't expecting a network graph question at the phone interview level. But guess it's facebook :)

    Interview Questions

    • To create a network graph from a log file in 20 mins. Guess the interviewer wanted a simple dictionary as a data structure but confused me with calling it a network graph repeatedly. If he indeed wanted it like a graph where all properties such as pre order , in order traversal could be applied. It's no way a 20 min question.   3 Answers

Don't Miss Out On a Job You Love
Upload a resume to easily apply to jobs from anywhere. It's simple to set up.