LinkedIn Data Scientist Interview Questions

LinkedIn·Data Scientist·Technical Phone Screen

Jul 2026

LinkedIn data scientist interview with a meaty A/B testing case built around Simpson's paradox. One question, but it had a lot of layers and the follow-ups kept coming.

A marketing team ran an A/B test on two email versions across two cities and two weeks. The pooled data shows version A winning, but version B wins in every city-level and week-level breakdown. How do you figure out which email is actually better? Walk through your choice of primary and guardrail metrics, how Simpson's paradox shows up here, what confounders you'd investigate, how you'd reanalyze the data, whether confidence intervals are still valid, and how you'd redesign the experiment if the original allocation was imbalanced.

“This one sprawled in a way I wasn't ready for.”

View Post

LinkedIn·Data Scientist·Technical Phone Screen

Jul 2026

LinkedIn DS interview with a SQL-heavy technical screen. Two parts to the same problem, nothing too exotic, but the histogram piece tripped me up more than I expected.

Given a table of article view events and a table mapping articles to their types, write SQL to find the number of distinct article types each user viewed on a specific date (2019-01-01).
Using the same tables but across all dates, build a histogram showing how many users viewed exactly N distinct article types. Return each possible value of N and the count of users at that value.

“Pretty straightforward join and group-by.”

View Post

LinkedIn·Data Scientist·Technical Phone Screen

Jul 2026

LinkedIn data scientist interview that went deep into probability theory, specifically around coin flip stopping times. Not your typical SQL or metrics question, this one felt more like a grad school exam than a product interview.

For a fair coin, what is the expected number of flips until you see two consecutive heads? Derive it using a state-based or Markov chain argument.
Now generalize: if the coin has probability p of heads, what is the expected number of flips to see HH?
For the fair coin case, derive the variance of T, the number of flips until HH appears. Walk through your full reasoning.

“I set up three states: start, one head seen, done.”

View Post

LinkedIn·Data Scientist·Onsite - Product Sense / Strategy

Jul 2026

Product data science interview at LinkedIn with two meaty product sense questions. Nothing behavioral, just pure case work. The kind of interview where you either have a framework in your head or you're winging it.

LinkedIn sees a material drop in weekly traffic from the Home Page to the Profile Page. How would you investigate whether this is a product bug, a logging issue, a user behavior shift, or an intentional improvement?
Design an experiment and success measurement framework for changing LinkedIn's Home Feed default from showing all content to showing only the most relevant content per viewer.

“The inline profile preview context is what makes this question interesting rather than generic.”

View Post

LinkedIn·Data Scientist·Technical Phone Screen

Jul 2026

LinkedIn data scientist interview focused almost entirely on a single product analytics case about diagnosing a drop in job applications. It was one of those sessions where you think you've covered everything and then the follow-up goes three layers deeper than you expected.

The number of completed job applications submitted per day on LinkedIn has dropped sharply. Walk through how you would systematically diagnose this, including the analyses you'd run, KPIs you'd track, visualizations you'd build, and data you'd request.
How would you separate a product regression from an external cause like seasonality, a macro labor market shift, or an SEO drop?
What segmentation cuts would you prioritize to localize where the drop is coming from?
What specific visualizations would you build to communicate your findings, and what data would you actually request to support them?

“This is the whole interview, basically.”

View Post

LinkedIn·Data Scientist·Technical Phone Screen

Jul 2026

SQL-heavy technical screen for a Data Scientist role at LinkedIn. Three progressively harder questions all on the same table, starting with a basic group-by and ending with window functions and percentage calculations. Felt manageable but the last part tripped me up a bit.

Given a table of countries with their continent and population, write a query to return the most populous country in each continent.
Follow-up: modify your solution to return the top 2 most populous countries per continent, including a rank column where 1 means the largest.
For those top 2 countries per continent, compute each country's share of the continent's total population, rounded to 2 decimal places.

“Pretty standard warm-up.”

View Post

LinkedIn·Data Scientist·Onsite - Behavioral / Leadership

Jul 2026

LinkedIn senior/tech-lead data scientist loop, all behavioral. Four prompts, all requiring real stories with real numbers. The whole thing felt like a deep audit of your resume more than a conversation.

Tell me about a time you led or mentored a team through an ambiguous project. How did you set direction, divide up the work, track progress, handle disagreements, and support people with less experience?
Describe a time you improved the quality of a product or model. How did you define what 'quality' even meant, pick the right metrics, find the root cause of problems, decide what to fix first, and verify it actually worked?
Walk me through a project you shipped that had clear, measurable business impact. How did you get stakeholders aligned, figure out the right scope to start with, navigate trade-offs, and demonstrate impact after launch?
Deep dive on your resume: specifically around what you personally owned, the decisions you made, how you influenced people outside your team, and what results you can directly attribute to your work.

“This one is deceptively broad.”

View Post

LinkedIn·Data Scientist·Onsite - Product Sense / Strategy

Jun 2026

LinkedIn data scientist interview focused entirely on a B2B product case, four questions all circling the same scenario. Heavy on product analytics and strategy thinking, less about coding.

A new B2B product just launched at LinkedIn. How would you build a metric framework to define what success looks like for it?
If you're looking at metric distributions and you notice metric A is up while metric B is down, how do you investigate that?
How would you determine whether this new product is actually delivering real value to its users, not just surface engagement?
Walk me through how you'd use data to identify and prioritize growth opportunities for this product going forward.

“I went with the classic funnel breakdown: acquisition, activation, engagement, retention, revenue.”

View Post

LinkedIn·Data Scientist·Technical Phone Screen

Jun 2026

LinkedIn data science interview with a stats-heavy question about user behavior on a video platform. Single question but they wanted the full treatment: data slicing, metric definition, test selection, and code. Not a vibe check, they actually wanted to see you work through it.

A video platform wants to know if users' first uploaded videos tend to be shorter than their later uploads. How would you test this hypothesis? Walk through the data slice, the metric, the statistical test you'd use, and sketch out the SQL or pseudo-code.

“I went straight for a paired t-test comparing first video duration vs.”

View Post

LinkedIn·Data Scientist·Technical Phone Screen

Jun 2026

LinkedIn data scientist interview with a bot detection problem that required both SQL and Python. Pretty applied, which I liked, but the open-ended definition of 'bot' left me second-guessing myself the whole time.

Given a table of page view events with member IDs and timestamps, write SQL and Python to identify and remove or flag bot users. You get to define what counts as a bot, so explain your reasoning.

“The open definition is what tripped me up at first.”

View Post