Meta Interview Questions — Page 36

Meta·Data Scientist·Technical Phone Screen

Apr 2026

Meta data scientist technical screen, basically one big SQL question that took up the whole session. The problem was a staggered-adoption diff-in-diff setup and it had enough moving parts that I was sweating by the end.

Given a schema with sites, employees, shuttle service adoption dates, and daily participation records, write SQL to build an individual-day panel for a staggered-adoption difference-in-differences analysis. The output needs employee_id, site_id, date, participated, the site-level adoption date, a treated indicator (1 if on or after adoption date, else 0, with never-treated sites getting NULL adoption and 0), event_time in days relative to adoption, and a post binary. Then also produce a weekly site-level aggregation with average participation rate, correctly handling never-treated sites.

“This wrecked me a little.”

View Post

Meta·Product Manager·Onsite - Product Sense / Strategy

Apr 2026

PM product sense question for Meta, framed as a Netflix scenario. Single question, no behavioral stuff, just pure product design and metrics thinking.

You're a PM at Netflix. Design a podcast feature for the platform: make the case for why it should exist, how you'd measure success, and which metrics you'd track.

“The 'why build it' part felt straightforward at first but I kind of fumbled the justification.”

View Post

Meta·Software Engineer·Technical Phone Screen

Apr 2026

Meta infrastructure engineer round, basically one coding problem the whole time. The problem itself wasn't hard conceptually but I fumbled the implementation more than I'd like to admit, and the follow-up discussion about memory tradeoffs was where things got interesting.

Given an immutable linked list where you can only call printValue() and getNext(), print all node values in reverse order. Then walk through multiple approaches with different time/space tradeoffs.

“I knew recursion immediately and coded it up fine.”

View Post

Meta·Data Scientist·Onsite - Behavioral / Leadership

Apr 2026

Behavioral round for a Data Scientist role at Meta. Four questions, all pretty standard for this type of interview, though the stakeholder pushback one tripped me up more than I expected.

Tell me about a time you had to pivot a project mid-way. What caused the pivot and what happened as a result?
Describe a situation where you pushed back on a stakeholder. How did you handle it?
Give an example of a time you gave constructive feedback to a teammate.
How do you approach picking up a completely new technical skill when you're under time pressure?

“I had a decent story for this one but I rambled too much on the context and ran short on the outcome.”

View Post

Meta·Software Engineer·Technical Phone Screen

Apr 2026

Meta SWE coding round, one meaty string problem that looks approachable until you realize the brute force is going to time out on anything real. The O(n) insight took me an embarrassingly long time to land on.

Given a string, compute the sum over all substrings of the count of characters that appear exactly once in that substring. How do you do this efficiently?
What are the edge cases and correctness concerns for the character-contribution approach, and how does it relate to the classic longest substring with all unique characters problem?

“Started with the obvious O(n^3) brute force and they let me run with it for maybe 90 seconds before asking if I could do better.”

View Post

Meta·Machine Learning Engineer·Technical Phone Screen

Apr 2026

Meta MLE screen with one algorithmic question that sounds easy but has a subtle implementation detail that'll trip you up if you've never thought carefully about weighted random sampling before.

Given an array of weights, implement a function that randomly returns an index where the probability of picking each index is proportional to its weight.

“I knew prefix sums were involved but fumbled the boundary condition for like two minutes.”

View Post

Meta·Software Engineer·Technical Phone Screen

Apr 2026

Meta SWE coding round, one algorithmic problem for the whole session. The problem looked deceptively approachable but had some edge cases that tripped me up mid-way through.

Given an integer array and a value k, find the length of the shortest contiguous subarray whose max minus min is at least k. Return -1 if no such subarray exists.

“My first instinct was brute force and I actually coded it up before catching myself.”

View Post

Meta·Software Engineer·Technical Phone Screen

Apr 2026

Meta infrastructure engineer screen, basically one algorithm problem the whole time. Binary search on an array to find a peak element. Seemed straightforward until I started second-guessing my edge case handling out loud.

Given an integer array where no two adjacent elements are equal, find any peak element (strictly greater than both neighbors) in O(log n) time. Boundary elements are considered to have negative infinity as their out-of-bounds neighbor.

“I knew binary search was the answer pretty fast, the O(log n) hint basically gives it away.”

View Post

Meta·Data Scientist·Technical Phone Screen

Apr 2026

Meta DS interview that was essentially a full research design case study. One big open-ended question about building a research plan for a new product feature, split into two parts. Felt more like a take-home prompt delivered live, which I was not expecting.

Before launching a group calling feature, how would you design a survey study to validate user needs? Walk through the specific questions you'd ask, the response formats, how you'd recruit participants, and how you'd connect survey responses to actual behavioral data without compromising privacy.
Without access to internal data, how would you estimate the U.S. adoption rate of a competitor's group calling feature to within roughly 20% relative error? What signals would you use, how would you collect them, and how would you triangulate them into a single estimate?

“This is where I spent most of my time and honestly undersold the vignette piece.”

View Post

Meta·Data Scientist·Technical Phone Screen

Apr 2026

A Meta DS interview that was basically a full product analytics case study compressed into one question. The depth required was pretty intense, covering everything from metric definition to experiment design to ship decisions.

How would you define and measure success for a goal of increasing the number of posts that receive at least one meaningful comment?
How would you establish a baseline and minimum detectable effect for this metric using recent data, and what seasonality concerns would you flag?
Generate at least 10 ideas to increase comment rates, spanning making it easier to comment, increasing intent to comment, better matching posts to likely commenters, and notification or feed changes. What are the risks and expected effect sizes?
What data would you collect to support analysis and experimentation for this feature?
Pick your top two ideas and design full experiments: unit of randomization, how you'd handle network interference, ramp plan, power assumptions, guardrails, and spillover diagnostics.
How would you attribute the lift from the experiment, distinguish incremental commenters from shifted commenting behavior, monitor for abuse or quality degradation, and make a ship or no-ship decision?

“I started with the obvious, proportion of new posts getting a non-deleted comment within 24 hours, but then realized I was glossing over what 'meaningful' actually means operationally.”

View Post