Are you preparing for the Facebook data scientist interview?

Hi, I’m Dan — a data scientist, previously at PayPal, now at Google. As an interview coach at datainterview.com, I want to help a candidate such as yourself ace data science interviews and land your dream role at a top company.

In this article, I will cover everything you need to know for your upcoming interview.

For more premium content, make sure to check out my prep site @ datainterview.com.

The data scientist position at Facebook is no doubt renowned and lucrative. According to Levels.fyi, the typical ranges of total compensation (TC) for a data scientist role at Facebook are:

*The compensation is based on Menlo Park, CA

Strictly speaking on salary alone, the pay at Facebook is no doubt higher compared to that of data scientists nationwide, which is about $123K. It’s no wonder that many aspiring data scientists dreaming of working at Facebook.

Please note the following caveats about this guide:

1. The process and preparation covered in this article are based on the product data scientist role which is one of the most sought-out roles at Facebook. For other roles such as Finance, Infrastructure, and Machine Learning, stay tuned at datainterview.com for comprehensive guides such as this one!
2. The guide doesn’t cover how to apply for a role and land the initial interview at Facebook. Tips on resume and cover letter will be covered in the near future.
3. The interview process at Facebook is constantly evolving. Note that the process and techniques covered here are based on interviews conducted in the past. Interviews in the future may be subject to change.

The Interview Process

In the Facebook DS interview, a candidate is funneled through a three-stage process:

For each interview step, we will cover the purpose, process, sample questions, and useful tips.

Stage 1 — Recruiter Screening

The first call with Facebook is a non-technical meeting with a recruiter. The call itself is about 20 to 30 minutes, and it is designed for recruiters to screen whether you are a decent fit for the role you applied for.

Before the call

The recruiter sees your job application that includes a resume and an optional cover letter in the applicant tracking system (ATS). Your application is algorithmically ranked based on how well your candidacy matches the roles described in the job posts. Recruiters will typically prioritize applications with higher ranking to contact first.

During the call

During the call, which is about 20 to 30 minutes, the recruiter will format the meeting in the following structure:

1. Details about the open role and Facebook’s mission — The recruiter will describe more details about the role expectation and team.
2. Candidate background — This is your chance to sell yourself verbally. The recruiter will ask, “Tell me about yourself.” You can provide a high-level description of your academic and career backgrounds. Some follow-up questions include: “Why do you want to work for Facebook?”
3. Technical screening with basic SQL question — As the first line of defense filtering out candidates who lack technical competencies, recruiters will ask basic SQL questions such as, “Can you explain the difference between INNER, LEFT, and OUTER JOINS?” And, “What’s the difference between UNION and UNION ALL?”
4. Logistics — The recruiter usually asks the following: Where are you located? Are you a U.S. citizen? If not, do you need an employer sponsorship for your visa? What are your availabilities for technical interviews?
5. Follow-Ups — The recruiter will detail the next steps in terms of when you should hear back and technical rounds. This is your chance to ask as many questions as you can to map out the technical interviews end-to-end. The more information you have, the more you can leverage it to prepare for interviews.

After the call

After the call, the recruiter will follow up with the hiring manager with notes gathered about the candidate’s background, technical screening, logistics, and culture fit. If the recruiter and hiring manager believe that you have potential, then they will advance you to the first technical round.

Preparation Tips

To demonstrate a really good impression, make sure you prepare the following:

1. Create a short elevator pitch explaining why you want to work for Facebook.
2. Project a friendly and positive impression during the call.
3. Brush up on SQL basics.
4. Prepare questions to ask in advance. For instance, ask questions that will help you gather as much information about the interviews as possible: how many rounds? What is the type of each round? Who is the interviewer? Gathering this information can help you design a prep strategy.

Stage 2 — Technical Phone Screening

The first technical screen is designed to test your technical aptitude in 45 minutes. Should you pass this round, you will proceed to the onsite rounds stage.

The technical screen for the product data science role is usually 45-minutes consisting of two question types covered:

1. SQL (15–20 minutes) — You will be given access to a non-executable text editor on the Coderpad and asked to solve two to three SQL problems involved in table manipulations. Typically they ask problems that involve table joins, distinct clauses, window functions, and where condition.
2. Product-Sense (15–20 minutes) — You will be given three to five open-ended questions involving metrics, analytics, and AB testing.

For sample questions, check out datainterview.com.

Preparation Tips

In preparation for the interview, do the following:

1. Practice mock interviewing with a friend or a fellow candidate when you join the datainterview.com Slack study group.
2. Be in the mode of interviewing with as many companies as possible. The more you interview, the less you will feel nervous and perform better over time.
3. Practice solving SQL problems on a daily basis.
4. To cultivate product-sense requires two steps: (1) become familiar with the Facebook core products and(2) get practice questions and see how data scientists at FAANG companies would solve them on datainterview.com.

Stage 3 — Onsite Rounds

The final technical screening at Facebook consists of four onsite (or virtual) rounds, each round consisting of 45-minutes. This is the longest and hardest stage of the interview process, and it will test your ability to sustain focus and solve problems under pressure. Let’s look at the details behind how the interview panel will assess you in each round:

SQL Round — Similar to the ones asked in the first technical round, the SQL questions asked in this round will require you to write an SQL logic on a non-executable text editor. Once you make sense of the problem, you will need to write and explain your solution. The interviewer will assess you based on your completion, solution efficiency, and communication.

Here are some questions Facebook asked:

“Given a table of user profile data and post metadata, what percent of users post within a month of their birthdays?”

“Given a table with user id and post id, calculate a 30-day moving median of the post count per user on each date.”

Applied Data Round — The applied data is a business case round designed to assess how you would solve an open-ended data problem. The interview questions are inspired by real problems that product data scientists work on given business stakeholder’s requests.

Here are sample questions:

“How do you know if two friends are best friends?”

How do you detect fake posts?

Quantitative Round — The quantitative is designed to assess your statistical and probability sense. The problems are designed to test your fundamentals on topics such as:

1. Bayes Theorem
2. Conditional Probability
3. Hypothesis Testing
4. P-Value
5. Confidence Interval
6. Basic Distributions (i.e. Bernoulli, Binomial, Normal)

Product-Sense Round — The product-sense round at Facebook is designed to screen your technical skills on AB testing, metric-sense, and product analytics. Here are sample questions:

Product Analytics: “A success metric increased in country A, but decreased in country B. Why?”

Metric-Sense: “How would you measure success on Instagram Stories?”

AB Testing: “How would you design experimentation that tests whether a new News Feed the algorithm improves engagement among users?”

Preparation Tips

In preparation for the interview, understand the following:

1. Practice daily by using datainterview.com to get access to Facebook interview questions and solutions.
2. To prepare for product-sense questions, become familiar
3. Practice mock interviewing with a friend or check out datainterview.com for coaching services that will launch in the near future!
4. Be in the mode of interviewing with as many companies as possible. The more you interview, the less you will feel nervous and perform better over time.
5. Practice solving SQL problems on a daily basis.
6. To cultivate product-sense requires two steps: (1) become familiar with the Facebook core products (2) get practice questions and see how experts would solve them on datainterview.com
7. The onsite interview will test your ability to sustain focus for a long period of time. Once a week, spend two to three hours practicing interview questions in one sitting. This will help build your stamina.

Research, Research, Research!

This part of the preparation is perhaps the most underrated step in acing a data science interview, especially at Facebook. From recruiter to onsite rounds, interviewers are assessing whether you have a sense of Facebook’s platform and its core products/features.

Just practicing problems isn’t enough. As a precursor, you do need to become somewhat familiar with Facebook’s products. When interviewers ask about case problems in product analytics and applied data rounds, you will thank yourself for the research you had done prior.

Research Tip #1 — Become a User

Become familiar with both the mobile and web versions of Facebook’s popular products:

1. Instagram
2. Facebook News Feed
3. Stories
4. Messanger
5. Groups
6. Posts
7. Profile

Spend some time making observations on how the product works and think about what kind of data would be collected in each of the products. This will help you if you were asked an open-ended question such as “How would you detect spammers? What kind of data would you use?”

Research Tip #2 — Keep Up with the Latest News

Also, another great way to familiarize yourself with Facebook is to watch their IO talks and read the latest articles about what Facebook is up to. This provides insights into what the business as a whole cares about. When you are asked business case problems, you can tailor your responses based on current issues. This shows that you’ve done your homework, which is something that the interviewer will take a positive note of.

Research Tip #3 — Read Facebook’s Engineering Blog

Lastly, read Facebook’s engineering blog which contains articles on the research and analysis that Facebook conducted. The blog offers clues on, not only interview questions, but also potential projects you could be working on.

Let’s Practice! 😃

Now, here’s the fun part of the article. Let’s practice with a sample interview question asked in a product-sense round by Facebook.

Facebook recently launched Stories, how would you measure success?

Bad Example

[Candidate] I’ve used Stories before. It’s a product that allows users to publish short-clip videos with a duration of 20 seconds at most. I think there are a couple of metrics that we can track.

[Interviewer] Okay, what are those?

[Candidate] I think metrics such as the number of watches per user and number of reactions per user are good.

[Interviewer] Can you think of anything else?

[Candidate] I think those are sufficient.

Interviewer Assessment

There are a couple of things that the candidate failed to do:

1. Failed to discuss the business problem in detail. The candidate should start by discussing what “success” means for Facebook.
2. The candidate failed to regard the audience in mind. Depending on the type of the stakeholder (analyst, product manager, marketer and e.t.c.), the types of metrics relevant will differ.
3. The candidate designed the metrics poorly as “number of watches per user” is a distribution, but not a statistic such as “average number of watches across users per day.”
4. The metrics are not ideal proxies for representing the “success” of the product.
5. The story is a two-sided app involving watchers and publishers. The candidate only mentioned metrics involving watchers.

Good Example

[Candidate] Thank you for the question. I’d like to approach this open-ended question in the following structure. First, ask clarifying questions. Next, discuss the business context of the problem. Lastly, propose success metric(s). Is this procedure okay with you?

[Interviewer] Sure.

[Candidate] When you are asking for measuring “success,” are you looking for a single metric that serves as a proxy that represents “success?”

[Interviewer] Yes.

[Candidate] Thank you. One more question. Who is the end-user of the metric? I would imagine the metric is displayed on a dashboard.

[Interviewer] Good question. Let’s say that the end-users are the Facebook leadership.

[Candidate] Great, thank you for the clarification. I can tell that Facebook cares a lot about the success of Stories as it’s a core product on the platform. It’s probably what keeps the users engaged in the long-term, and one of the top revenue-generating products as advertisements are displayed on Stories.

[Interviewer] That’s correct.

[Candidate] Now, that I have some sense of the business problem. I’d like to propose three metrics — a single primary metric that the leadership can say, “we want to increase this driver metric by 10% YoY” and two secondary metrics to supplement the information that the primary metric alone does not provide.

[Interviewer] Great, proceed.

[Candidate] The primary metric should be total watch hours per month. I can imagine the stakeholder would want to understand how the Stories grew over a long period of time, and they want to get a sense of how this will be projected over time. So, looking at the total watch hours at the monthly level can provide information on how the product as a whole grew.

In terms of secondary metrics, we also want to track publisher behavior. So, looking at the total number of videos published per month is a good metric. On a long-term, month-to-month basis, we want to see this metric grow. More content available means more videos that watchers can select and consume.

Another secondary metric involves the advertisement revenue of the Story product. We want to see that this product is contributing to Facebook’s bottomline.

Interviewer Assessment

The candidate provided a superb response:

1. The candidate provided a structured response: First, ask clarifying questions. Next, discuss the business context of the problem. Lastly, propose success metric(s).
2. Understood the business problem and audience.
3. Provided metrics that were well-designed based on the business problem.
4. Communicated the solution, eloquently.

Next Step!

Do you want more practice questions? Make sure to check out datainterview.com!