← Back to Directory

Meta interview notes, page 34 of 238

FAANG / Big Tech

Most of these Meta write-ups come from Software Engineers (1,194), then Data Scientists (489) and Product Managers (453). The most reported rounds are the technical phone screen (1,190), the product sense round (466), the system design round (255) and the behavioral / leadership round (205).

2377 interview notes · updated Jul 2026

Meta·Machine Learning Engineer·Technical Phone Screen

Apr 2026
Verbal technical screen for an ML Engineer role at Meta, focused entirely on transformer fundamentals. Three main topics came up in sequence and the interviewer seemed to want derivations, not just high-level answers.
  • Derive the scaled dot-product attention formula and explain why we divide by the square root of the key dimension.
  • Walk through every component inside a single transformer block, and explain the difference between pre-norm and post-norm placement.
  • Compare causal and bidirectional attention masking. Where is each actually used?
  • Compare different positional encoding schemes, including sinusoidal, learned absolute, relative bias approaches, and rotary encodings. What are the tradeoffs around extrapolating to longer sequences?
  • How does dropout behave differently during training versus inference, and why do we scale activations the way we do?

“I could write out softmax(QK^T / sqrt(d_k))V fine but stumbled explaining the scaling part.”

View Post

Meta·Software Engineer·Technical Phone Screen

Apr 2026
Interviewed for a data science role at Meta and got hit with a product analytics question about fake news that was deceptively open-ended. Not a lot of structure to the original prompt so I'm piecing this together from what I remember.
  • How would you measure and evaluate the impact of fake news on Facebook users?

“This question is way bigger than it looks.”

View Post

Meta·Software Engineer·Technical Phone Screen

Apr 2026
Meta SWE coding round with a string manipulation problem that looks like a word break variant but has a suffix-specific twist. Not a lot of context shared about how it went, just the problem itself.
  • Given a list of words and a list of target strings, determine for each target whether it can be formed by concatenating one or more words from the list (words can be reused). Return a boolean list.

“Classic word break problem dressed up with a 'suffix' label, which threw me for a second because I kept thinking it meant something about actual string suffixes.”

View Post

Meta·Product Manager·Onsite - Product Sense / Strategy

Apr 2026
Meta PM interview with a product design question about building something for gardeners. Pretty open-ended, which I wasn't fully prepared for.
  • Design a product for gardeners.

“I spent the first minute just trying to narrow the user group because 'gardeners' is so broad.”

View Post

Meta·Data Scientist·Technical Phone Screen

Apr 2026
Meta Data Scientist technical screen, heavy SQL focus. The whole thing was basically one long multi-part query problem built around survey response rates and quality scoring. Felt like a take-home but live.
  • Given an impressions table, a clicks table, a surveys table, and a survey_responses table, write SQL to compute the daily survey response rate (unique clicked impressions divided by total impressions) broken out by survey version, for a 7-day window ending today.
  • Compute the overall response rate for the full 7-day window as a single number, and then also break it out by survey version.
  • For the same 7-day window, compute two quality metrics by survey version: one using only each user's chronologically first score per survey, and one averaging all scores including duplicates from the same user.
  • Modify your query so that a survey response is only counted if there is a corresponding impression for that user and survey within the same 7-day window, and the response timestamp is at or after the first click for that user and survey on that day. Responses with no matching impression should be excluded.
  • Combine all of the above into a single result set returning: date, survey version, total impressions, unique clicked impressions, response rate, average score using first-only method, and average score using all scores. State any assumptions you make about timezone and missing joins.

“The join to the surveys table to pull in version was the easy part.”

View Post

Meta·Software Engineer·Technical Phone Screen

Apr 2026
Meta Data Engineer coding screen, one question the whole time. Pretty algorithmic for a DE role but not unreasonable once you see what they're actually asking for.
  • Given a list of books where each book has a category and a point value, return the maximum total points a student can earn by selecting at most 3 books, with the constraint that no two selected books can share the same category.

“My first instinct was to reach for combinations and just brute force it, which would've been fine for small inputs but they flagged upfront that the list could be large.”

View Post

Meta·Software Engineer·Online Assessment (OA)

Apr 2026
Meta SWE coding round, four back-to-back problems covering a pretty wide range of topics. Nothing felt impossible but the matrix transformation one slowed me down more than I expected.
  • Given two equal-length arrays of ratings and prices, find the index with the highest ratings-to-price ratio. Ties go to the smaller index, and you must compare fractions exactly without floating-point division.
  • You need to complete a given number of round trips between two locations using sorted departure time arrays. Each leg requires picking the earliest available time at or after your current time. Return the finish time or -1 if it's impossible.
  • Execute a sequence of matrix transformation commands in order: swap rows, swap columns, reverse a row, reverse a column, or rotate the whole matrix 90 degrees clockwise. Return the final matrix state.
  • Count the number of ordered index pairs (i, j) in an array of strings where concatenating fragments[i] and fragments[j] equals a target string. Duplicates count separately, and i can equal j.

“The no-floating-point constraint is the whole point of this problem and I almost missed it.”

View Post

Meta·AI Engineer·Technical Phone Screen

Apr 2026
Meta AI Engineer coding screen, just one question but they really sat with it. Sliding window stuff, classic but easy to fumble under pressure.
  • Given a string, find the length of the longest substring where all characters are unique.

“Knew the sliding window approach going in, which helped.”

View Post

Meta·Software Engineer·Technical Phone Screen

Apr 2026
Meta SWE coding round, pretty much a string filtering problem. Nothing crazy but it's the kind of thing where you can overthink it if you're not careful.
  • Given an array of strings and a set of characters, return all strings from the array that can be formed using only characters from that set.

“Straightforward on the surface.”

View Post

Meta·Software Engineer·Technical Phone Screen

Apr 2026
Meta SWE coding round, one question the whole time. It looked like a stock trading problem but the cooldown twist is what they actually care about. Took me a minute to realize brute force wasn't going to cut it.
  • Given an array of daily stock prices, find the maximum profit you can make from unlimited buy/sell transactions, with the constraint that after selling you must wait one day before buying again.

“I knew the basic buy-low-sell-high problem cold, so I started heading down that path and had to stop myself.”

View Post