← Back to Directory

Meta interview notes, page 229 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·Data Scientist·Technical Phone Screen

Jul 2026
Meta DS interview focused entirely on a product case around launching a video-calling feature inside a messaging app. Four connected sub-questions that built on each other, which I wasn't fully prepared for. The metric trade-off question at the end was the one that really made me sweat.
  • Using only a provided SQL dataset, how would you determine whether users actually need or want a new video-calling feature before it launches?
  • If you had access to all internal data sources beyond just SQL, what additional analyses would you run to measure demand for the video-calling feature?
  • After the video-calling feature is released, how would you evaluate its quality and overall success?
  • If your primary success metric improves but a guardrail metric declines, how do you proceed and how do you communicate that decision to stakeholders?

“I started talking about usage frequency and message volume as proxies for engagement, then pivoted to looking at things like group chat size or call-adjacent behaviors in the data.”

View Post

Meta·Machine Learning Engineer·Technical Phone Screen

Jul 2026
Meta MLE coding round, two algorithm problems back to back. Nothing behavioral, just pure coding. The second one was harder than I expected for the role.
  • Given a nested list that can contain integers or further nested lists, compute the weighted sum where each integer is multiplied by its depth in the nesting (top-level elements start at depth 1).
  • On an m x n grid with buildings, empty land, and obstacles, find the empty cell whose total shortest-path distance to all buildings is minimized. Return that minimum distance, or -1 if no such cell exists.

“Recursive DFS, pretty clean once you see it.”

View Post

Meta·Software Engineer·Technical Phone Screen

Jul 2026
Meta SWE coding round with a classic array problem that has a nasty O(n) constraint attached to it. Straightforward premise, less straightforward execution.
  • Given an unsorted array of integers, find the smallest missing positive integer. Your solution must run in O(n) time and use constant extra space.

“The example they give looks simple enough, [3, 4, -1, 1] returns 2, sure fine.”

View Post

Meta·Software Engineer·Onsite - Coding / Algorithms

Jul 2026
Meta SWE coding round focused on a maze problem split into two staged tasks: fixing a buggy printer function and then implementing a BFS solver. The AI-assisted format meant you were dropped into an existing codebase rather than starting from scratch, which added a layer of 'read before you write' pressure I didn't fully appreciate until I was already in it.
  • You're given a buggy function that prints a 2D character grid representing a maze. The output comes out wrong (transposed, extra spaces, or extra blank lines). Find and fix the bug so the maze prints row by row with no extra whitespace.
  • Implement a BFS-based function that finds the shortest path length from 'S' to 'E' in a 2D maze grid, moving only in four orthogonal directions. Return the number of moves, or -1 if no path exists.
  • How would you reconstruct the actual path of cells visited, not just the length, and what changes in your bookkeeping to support that?
  • What would you change if cells had non-uniform movement costs, or if some cells were zero-cost teleports?
  • Could an A* heuristic (like Manhattan distance) improve performance here, and what does it actually change about the complexity?

“The transposed loop thing is one of those bugs that looks fine at a glance because grid[c][r] is syntactically valid.”

View Post

Meta·Software Engineer·Technical Phone Screen

Jul 2026
Meta SWE coding round, one question the whole time: longest consecutive sequence. Seemed straightforward until they pushed for the O(n) version and wanted me to justify the amortized complexity out loud.
  • Given an unsorted array of integers, find the length of the longest run of consecutive integers. Can you do it in O(n)?

“I jumped straight to the sort-and-scan approach because it felt clean, and the interviewer just nodded and said 'can we do better.' Classic.”

View Post

Meta·Software Engineer·Technical Phone Screen

Jul 2026
Meta SWE coding round, one problem the whole session. Clean premise but the edge cases sneak up on you if you're not careful.
  • Given a list of lowercase English words, find the largest subset where every character across all chosen words appears exactly once. Words with internal duplicate letters are disqualified. Return the total count of unique characters covered.

“My first instinct was greedy but that falls apart fast.”

View Post

Meta·Data Scientist·Technical Phone Screen

Jul 2026
Meta data scientist interview, looked like a single deep technical question that sprawled into like five sub-problems at once. The whole thing revolved around multi-armed bandits with real-world messiness layered on top. Pretty brutal if you're not already living in experimentation land.
  • You have three product variants, a churn guardrail metric, and delayed conversion signals. Design a Thompson Sampling bandit from scratch: specify the likelihood model, explain how you handle conversion delays (optimistic vs. debiased estimators), and address non-stationarity through discounting or similar approaches.
  • How would you set traffic floors and fairness constraints across the variants in this bandit setup?
  • Define stopping rules and rollback policies for when a guardrail metric gets breached during the bandit experiment.
  • Compare the expected regret and business impact of this bandit approach versus a fixed-horizon A/B test, specifically under seasonal conditions.

“This is where I spent most of my mental energy and still feel like I only half-answered it.”

View Post

Meta·Data Scientist·Technical Phone Screen

Jul 2026
Meta DS interview with a product scenario about ad targeting strategy. One meaty question that pushed into monetization tradeoffs, not just metrics.
  • A PM wants to serve ads exclusively to high-intent users. Walk through the pros and cons across revenue, user experience, and long-term growth.

“I jumped straight to 'more revenue per impression' and felt pretty good about that angle, but then kind of stalled when thinking about what you lose.”

View Post

Meta·Data Scientist·Technical Phone Screen

Jul 2026
Meta DS interview focused entirely on a push notification case study, walking through metrics, user complaints about overload, and then a fairly involved algorithm design problem. The whole thing felt like one long scenario with sub-questions rather than separate rounds, which I wasn't expecting.
  • What metrics would you use to measure the quality of push notifications?
  • Users are complaining that they're receiving too many push notifications. How would you address this?
  • How would you design a per-user daily push notification cap that maximises open rates without increasing uninstalls?

“Went straight to open rate and kind of stopped there for a beat too long.”

View Post

Meta·Data Scientist·Technical Phone Screen

Jul 2026
Meta DS interview focused on ML system design, specifically ranking posts for Newsfeed. The whole thing was one big open-ended problem and they clearly wanted to see how deep you could go on features, modeling choices, and evaluation.
  • Design a machine learning model to rank Facebook Groups posts in a user's Newsfeed. Walk through your feature choices, labels, model family, and how you'd evaluate and deploy it.

“This is a beast of a question and I underestimated how much they'd push on each layer.”

View Post