← Back to Directory

Meta interview notes, page 186 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·Software Engineer·Technical Phone Screen

Jun 2026
Two coding problems for a Meta SWE round. The maze one was manageable once I remembered to encode key state in the BFS node, but the suffix query problem had a performance requirement that made me rethink my whole approach mid-interview.
  • Given a 2D grid maze with keys and locked doors, find the minimum number of steps to travel from the start to the exit, where doors can only be passed if you've already collected the matching key.
  • You have a large list of words and a large list of queries. For each query, find the word with the longest common suffix, using word length and then index as tiebreakers. Brute force is too slow at scale, so implement something better and benchmark it.

“My first instinct was plain BFS and I coded half of it before realizing the state space is wrong.”

View Post

Meta·Machine Learning Engineer·Onsite - System Design / Architecture

Jun 2026
Meta ML Engineer system design round, basically one big question about recommender systems that kept branching into sub-topics for the whole session. Solid prep on two-tower models helped but I underestimated how deep they'd go on the serving and feedback loop side.
  • Design an end-to-end recommender system for a content or product feed, covering candidate retrieval, ranking, re-ranking, and online serving.
  • How would you construct training data for the ranker, and how do you handle delayed or implicit feedback signals?
  • How do you handle cold start for both new users and new items in a recommender system?
  • What metrics would you use to evaluate the recommender system offline versus online, and how do you decide when offline metrics are sufficient?
  • How would you design the online serving infrastructure to support real-time inference at high query volume, including feature consistency between training and serving?
  • How do you think about feedback loops and filter bubbles in a deployed recommender system, and what can you do to mitigate them?

“This was the whole interview, basically.”

View Post

Meta·Machine Learning Engineer·Technical Phone Screen

Jun 2026
Meta ML engineer interview with a coding problem that was more algorithmic than I expected. No ML theory, just pure backtracking. Felt like a software engineering round with an ML job title on it.
  • Given a table of cards with numeric values, implement a backtracking algorithm to find the combination of 3-card sets (each set summing to 15) that maximizes the total number of cards used.

“Took me a minute to even understand what they were optimizing for.”

View Post

Meta·Product Manager·Onsite - Product Sense / Strategy

Jun 2026
PM interview at Meta with a product metrics question framed around Zoom. Pretty standard case-style format but the cross-company angle threw me a bit.
  • You are a PM at Zoom. What goals and metrics would you track?

“The Zoom framing is a little weird when you're interviewing at Meta, took me a second to just accept it and treat it as a pure product thinking exercise.”

View Post

Meta·Software Engineer·Technical Phone Screen

Jun 2026
Meta data engineer SQL question, just one problem but it had enough layers to keep me busy for a while. The referrer logic tripped me up more than I expected.
  • Given a copies table (with copy_id and reserved_by_member_id) and a members table (with member_id and referred_by_member_id), find the member with the largest member_id. Return their member_id, referred_by_member_id, and a calculated column representing the difference between the number of copies they reserved and the number of copies their referrer reserved. If the referrer is NULL or has made no reservations, treat their reservation count as zero.

“I got the basic structure pretty quick, aggregate copies by member, join to members, grab the max member_id.”

View Post

Meta·Product Manager·Onsite - Product Sense / Strategy

Jun 2026
Got a product design question at Meta about building a movies feature on Facebook. Pretty open-ended, not a lot of detail to go on from the original prompt, but the kind of question that can go sideways fast if you don't anchor it early.
  • How would you design a movies product for Facebook?

“The scope here is massive and I think that's the whole point.”

View Post

Meta·Software Engineer·Technical Phone Screen

Jun 2026
Meta SWE coding round, one question the whole session. The problem sounds straightforward until you realize they want O(log n) and suddenly you're staring at binary search trying to convince yourself why it works.
  • Given an integer array where no two adjacent elements are equal, find any peak element. A peak is an index where the value is greater than both its neighbors, with the boundaries treated as negative infinity. Solve it in O(log n).

“I knew binary search was the answer pretty fast but explaining WHY it works took me longer than I'd like to admit.”

View Post

Meta·Software Engineer·Technical Phone Screen

Jun 2026
Meta phone screen for a SWE role that leaned heavily on tree traversal problems. The vertical order and right side view questions came up as a combo, which I half-expected but still fumbled parts of.
  • Given a binary tree, return the nodes visible from the right side (right side view).
  • Given a binary tree, return its vertical order traversal grouped by column.
  • Can you return both the left view and the right view of a binary tree in a single BFS pass?

“I went with BFS and grabbed the last node of each level, which worked fine.”

View Post

Meta·Data Scientist·Technical Phone Screen

Jun 2026
Meta DS interview with a meaty ML design question about fake news detection. Single question but they clearly wanted you to go broad and then deep, covering features, modeling, and ops. Felt more like a system design round than a typical ML theory quiz.
  • How would you design a machine learning system to automatically detect and flag fake news on Facebook? Walk through the features you'd use, your modeling choices, how you'd evaluate it, and how you'd handle the model degrading over time.

“This question is deceptively wide.”

View Post

Meta·Data Scientist·Technical Phone Screen

Jun 2026
Meta DS interview focused entirely on product experimentation for a short-video feed context. Two meaty questions back to back, no warmup, no behavioral fluff. Felt more like a product analytics case than a traditional data science screen.
  • What are the top three metrics you would use to measure user engagement in a short-video feed?
  • Design an A/B test for a new recommendation algorithm: walk through how you'd calculate sample size, decide on experiment duration, and define stopping criteria.

“I went with watch time per daily active user, like-to-view ratio, and session frequency.”

View Post