Meta Interview Questions — Page 216

Meta·Machine Learning Engineer·Technical Phone Screen

Jul 2026

Meta MLE interview with three back-to-back coding and design questions, all leaning heavier on algorithms than I expected for an ML role. The problems were well-scoped but the follow-ups got tricky fast.

Given an m x n integer matrix, check whether every top-left to bottom-right diagonal contains identical values (the Toeplitz property). Return true or false, analyze time and space complexity, and explain how you'd handle the case where rows arrive one at a time as a stream.
Design a data structure that tracks the moving average of the last k values from a real-time number stream. It should support push and query operations in amortized O(1) time and O(k) space. Also address numerical precision, potential overflow, and what happens when fewer than k elements have been pushed.
Given an array and a window size k, return the median of each sliding window as it moves across the array. Aim for O(n log k) time or better. Walk through your data structure choices, how you handle duplicates and even-length windows, and discuss any memory trade-offs.

“The base matrix check was fine, just compare each cell to its upper-left neighbor and you're done in O(m*n) time.”

View Post

Meta·Software Engineer·Onsite - Behavioral / Leadership

Jul 2026

Meta staff-level behavioral round, one question deep. The whole session was basically a single mentorship prompt and they expected a lot of specificity about scope and measurable impact.

Tell me about a time you meaningfully mentored another engineer, such as helping someone grow into a more senior role or transferring ownership of a critical area to them.

“The part that tripped me up was the implicit bar they set.”

View Post

Meta·Software Engineer·Onsite - Coding / Algorithms

Jul 2026

Meta's AI-enabled coding round for a Software Engineer role. The problem was a classic container/word problem dressed up with a custom name, and the whole session revolved around optimization decisions and benchmarking two different approaches with AI assistance throughout.

Given a brute-force implementation of a word container problem, how would you identify the bottleneck and optimize it?
Compare the two solutions by integrating them into a benchmark and analyzing the results.

“They called it 'ParanoidEcho' which threw me for a second, but it was basically the word container problem.”

View Post

Meta·Software Engineer·Technical Phone Screen

Jul 2026

Meta SWE coding round, one meaty problem about building an expression evaluator from scratch. The operator precedence and right-associativity stuff made it trickier than it looked at first glance.

Implement a calculator that evaluates arithmetic expressions with non-negative integers and the operators +, *, and ^ (exponentiation), where ^ has highest precedence and is right-associative. Spaces may appear anywhere in the input. Return the integer result and discuss your approach.

“I went with recursive descent because two-stack shunting-yard always trips me up when I try to code it live.”

View Post

Meta·Data Scientist·Technical Phone Screen

Jul 2026

Meta DS interview focused entirely on a single multi-part case around Instagram shopping. The question was dense and covered market sizing, experiment design, and debugging all in one go, which felt like a lot to juggle in real time.

Estimate the annual revenue opportunity of launching a shopping feature on Instagram.
Design an A/B test to measure the impact of the Instagram shopping feature on GMV. Include your randomization unit, primary and guardrail metrics, and how you'd approach sample size.
Partway through the experiment, the treatment group's conversion rate drops sharply. Walk through how you'd troubleshoot this.

“I went TAM to adoption to take-rate, which felt right structurally, but I fumbled the take-rate number pretty badly.”

View Post

Meta·Software Engineer·Technical Phone Screen

Jul 2026

Meta SWE coding round with a probability/random sampling problem. The prefix sum approach is pretty standard but the boundary conditions tripped me up more than I expected.

Design a weighted random index picker: given an array of weights, implement an init method and a pickIndex method that returns an index with probability proportional to its weight, with pickIndex running in O(log n) time.

“The core idea clicked fast enough, build a prefix sum array and binary search on a random number in the total range.”

View Post

Meta·Software Engineer·Onsite - System Design / Architecture

Jul 2026

Meta data engineer interview, one system design question about building a warehouse schema for Instagram. Pretty open-ended and I wasn't totally sure how deep they wanted me to go.

Design a data warehouse schema for Instagram.

“I started with users and posts as the obvious fact/dimension split, but then spent way too long debating whether to normalize engagement events or just denormalize everything into wide tables.”

View Post

Meta·Product Manager·Onsite - Product Sense / Strategy

Jul 2026

Meta PM interview with a product strategy question about Facebook hashtags. Pretty open-ended, which I wasn't fully prepared for.

What goals would you set for Facebook hashtags?

“I went straight to engagement metrics and kind of forgot to first establish what problem hashtags are even solving on Facebook specifically.”

View Post

Meta·Software Engineer·Onsite - Coding / Algorithms

Jul 2026

Meta SWE coding round, one question that looks like an intersection problem but actually wants a union. Pretty quick once you read it carefully, though the mislabeling is a little annoying.

Given two integer arrays, return an array containing all unique elements from both arrays combined (their union), despite the problem being labeled as an intersection.

“The name threw me off for a second.”

View Post

Meta·Software Engineer·Technical Phone Screen

Jul 2026

Meta SWE screen that was basically a rapid-fire ML concepts quiz. Five questions, all conceptual, no coding. Felt more like a written exam than a conversation.

For a binary classification problem, define TP, FP, TN, and FN, then write out the formulas for accuracy, precision, recall, and F1. Also explain in plain English what precision and recall actually measure.
What is ensemble learning, why does combining multiple models tend to improve performance, and what are the common ways to aggregate model outputs?
Compare bagging and boosting: how each builds its training sets, whether each reduces bias or variance, the tradeoffs involved, and name a concrete algorithm for each.
Write the objective functions for L1 and L2 regularization, explain how each affects the learned weights differently, and discuss when you'd choose one over the other.
For a two-layer feedforward neural network, write the mathematical expressions for the hidden layer activations and the final output, then describe how you'd actually compute the output numerically given specific values.

“I know this cold so I wasn't worried, but I always trip myself up on whether F1 is optional or expected.”

View Post