← Back to Directory

Meta interview notes, page 101 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

May 2026
Meta coding screen, one problem the whole time. Pretty focused session, no fluff.
  • Given three sorted integer arrays, merge them into a single sorted array with all duplicates removed.

“My first instinct was to just concat everything and sort, which works but feels lazy for a Meta interview.”

View Post

Meta·Software Engineer·Technical Phone Screen

May 2026
Meta SWE interview with a tree problem that sounds straightforward until you realize the root isn't given to you. The parent pointer twist changes everything about how you approach it.
  • Given two nodes p and q in a binary tree where every node has a parent pointer, find their lowest common ancestor. You are not given the root directly and must navigate using parent and child pointers.

“My first instinct was to just walk both nodes up to the root and collect the paths, then compare.”

View Post

Meta·Software Engineer·Technical Phone Screen

May 2026
Meta Production Engineer interview with a coding round focused on a grid traversal problem. The follow-up discussion got surprisingly nitpicky about micro-optimizations, which I wasn't expecting at all.
  • Given a 2D board containing 'X' and '.', count the number of battleships. Battleships occupy 1×k or k×1 cells and never touch each other. Solve it in O(R·C) time with O(1) extra space.

“The key insight is you only count a cell if its top neighbor and left neighbor are both not 'X', meaning it's the top-left corner of a ship.”

View Post

Meta·Product Manager·Technical Phone Screen

May 2026
Went through what seemed like a product or strategy screen at Meta and got hit with a classic market sizing question. Not much else to go on but it was a decent brain-teaser.
  • How many mattresses are sold annually in the US?

“Market sizing.”

View Post

Meta·Product Manager·Onsite - Product Sense / Strategy

May 2026
PM interview at Meta focused on Instagram Reels monetization. Just one question but it went pretty deep into metrics and experimentation thinking.
  • You're the PM for Instagram Reels and you want to test new advertising formats. What metrics would you track?

“I started with engagement stuff like view-through rate and skip rate, which felt obvious, but then I got a bit tangled trying to separate advertiser-side metrics from user experience metrics.”

View Post

Meta·Product Manager·Onsite - Product Sense / Strategy

May 2026
Meta PM interview with a product design question around building a Facebook feature for blood donation. Pretty open-ended, which I wasn't fully ready for.
  • How would you design a Facebook product to facilitate blood donation?

“I jumped straight into features without anchoring on the problem first, which I could tell didn't land well.”

View Post

Meta·Data Scientist·Technical Phone Screen

May 2026
Meta DS interview focused almost entirely on experiment design for a shop ads ranking algorithm change. Five distinct sub-problems in one question, which felt like a lot to cover in a single session. The depth expected was real.
  • You're evaluating a new ranking algorithm for shop ads. How would you choose the randomization unit for the experiment, and what tradeoffs come with each option?
  • What primary and guardrail metrics would you use to evaluate whether the new ads ranking algorithm is working?
  • How would you approach a power analysis for this experiment, especially given that revenue metrics tend to be heavily skewed?
  • CTR goes up but revenue goes down after the algorithm change. Walk me through at least four reasons why that might happen and how you'd investigate.
  • What common pitfalls would you watch out for when running this kind of ads experiment, and how would you catch them?

“I went with user-level randomization pretty quickly and they pushed back asking why not session.”

View Post

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

May 2026
Meta ML Engineer system design round, one massive question about building an async RL post-training pipeline on top of a live serving system. The scope was brutal and I don't think I got close to covering everything they wanted.
  • Design an asynchronous RL-based post-training system (like RLHF or RLAIF) for a chat LLM that is already serving live traffic. Cover the major components such as generators, reward inference, learners, replay buffers, and orchestrators, how data flows between them, batching strategies, off-policy corrections like importance sampling or V-trace, and KL divergence control relative to the base model. Also address safety guardrails, versioning and rollout strategy, online feedback ingestion, credit assignment under delayed or sparse rewards, reward hacking prevention, separation of serving and training clusters, stability monitoring, and cost predictability under async feedback and load spikes.

“This was essentially a dissertation prompt compressed into 45 minutes.”

View Post

Meta·Software Engineer·Technical Phone Screen

May 2026
Meta SWE coding round, one algorithmic problem, pretty standard fare for that company.
  • Given an unsorted array of integers, find the kth largest element. The target is the kth largest in sorted order, not the kth distinct value, and you need to hit O(n) average time complexity.

“The O(n) constraint is the whole point of the problem.”

View Post

Meta·Software Engineer·Online Assessment (OA)

May 2026
Meta SWE online assessment, one problem focused on extending an in-memory record store to handle timestamps and TTL expiration. Pretty niche design problem, not your typical LeetCode grind.
  • Design and implement an in-memory record store that supports write timestamps and time-to-live (TTL) expiration. Records past their TTL should be invisible during reads and scans.

“This was a level 3 extension of a simpler key-value store problem, so you had to build on top of existing add/remove/scan logic and layer in expiration semantics.”

View Post