Meta
Meta Interview Process & AI-Enabled Coding Round (2026)
Updated every weekLast refreshed 9 Nov 2026Next refresh 16 Nov 2026
This page has 28 reports so far; 64% of those with a known location are from the US. The rounds they mention most: technical interview (24), system design (19), behavioral (18).
Meta runs a different software engineer loop at each level: E4 (Software Engineer) has 4 rounds; E5 (Senior Software Engineer) has 4 rounds. Each level below lists every round with its length and format, whether your code is run, and the AI policy.
Rehearse the Meta interview before the real one
The AI interviewer is set up for Meta. It asks a question, pushes on your answers with follow-ups, and grades you at the end, so you know what to fix.
Start a mock interview for MetaMeta interview process by level
Verified loops: every round with its length and format, and whether your code is run.
- Meta E4 (Software Engineer) · 4 rounds
- Meta E5 (Senior Software Engineer) · 4 rounds
Can you use AI?
Many Meta technical interviews now include an AI assistant built into the CoderPad environment, and candidates are expected to use it. No outside AI tools or assistance are authorized. Compare with other companies
Guides for the Meta interview
Recently reported Meta interview questions
Pulled from the reports below and updated every week. When a question matches a problem we have, you can practise it in the code editor.
Implement an LRU cache.
Phone screen · Coding · last seen 20 May 2026
Practise this problemMatches LRU Cache. Open it in the code editor and run it against test cases.
Breadth-first traversal of a binary tree.
Phone screen · Coding · last seen 20 May 2026
Practise this problemMatches Binary Tree Level Order Traversal. Open it in the code editor and run it against test cases.
Use BFS on a grid to reach the bottom-right cell and return the path taken.
Technical interview · Coding · last seen 20 May 2026
Practise this problemMatches Shortest Path in Binary Matrix. Open it in the code editor and run it against test cases.
Compute the sum of a nested list where each element is multiplied by its depth level.
Technical interview · Coding · last seen 20 May 2026
Design a Ticketmaster-style ticket purchase flow for a Taylor Swift concert with only 20k tickets, handling scalability and strong consistency.
System design · System design · last seen 20 May 2026
In an AI-assisted coding round, fix a grid-traversal program to reach the end with directional arrows and walls, passing as many of five tests as possible.
Other round · Coding · last seen 20 May 2026
Sum the numbers formed by root-to-leaf paths in a binary tree.
Phone screen · Coding · last seen 10 May 2026
Count palindromic substrings in a string, with requests to optimize space and reach the expand-around-center approach.
Phone screen · Coding · last seen 10 May 2026
Practise this problemMatches Palindromic Substrings. Open it in the code editor and run it against test cases.
Find buildings with an ocean view; follow-ups: ocean on both sides, and choosing one building to remove to maximize ocean-view buildings.
Technical interview · Coding · last seen 10 May 2026
Given M sorted arrays with N total elements, find the Kth smallest element overall.
Technical interview · Coding · last seen 10 May 2026
Compute the average of each sliding window of size k in an array in O(n) time and O(1) space, handling invalid inputs.
Technical interview · Coding · last seen 10 May 2026
Practise this problemMatches Maximum Average Subarray I. Open it in the code editor and run it against test cases.
Find a local minimum in an array in sub-linear time, adapting for strict comparison and duplicates.
Technical interview · Coding · last seen 10 May 2026
Practise this problemMatches Find Peak Element. Open it in the code editor and run it against test cases.
Design a crowdsourced 3D street view using smartphone video from cars, showing different renderings by season or time of day.
System design · System design · last seen 10 May 2026
Tell me about a time style behavioral questions with deep follow-ups.
Behavioral · Behavioral · last seen 10 May 2026
Fix and extend code to pass four test case files, using AI assistance for some parts and explaining the approach for one implementation.
Other round · Coding · last seen 16 Apr 2026
Solve a medium problem involving a binary tree using depth-first search, explaining brute force, optimal solution, complexity and a dry run.
Technical interview · Coding · last seen 16 Apr 2026
Solve a medium binary search problem, explaining brute force, optimal solution, complexity and a dry run.
Technical interview · Coding · last seen 16 Apr 2026
Design a system, defining functional and non-functional requirements, core entities, API endpoints and a high-level architecture.
System design · System design · last seen 16 Apr 2026
Merge overlapping intervals in a list; follow-ups on handling a real-time stream of intervals and duplicates.
Phone screen · Coding · last seen 9 Apr 2026
Practise this problemMatches Merge Intervals. Open it in the code editor and run it against test cases.
Find the lowest common ancestor of two nodes in a binary tree, including the case where both nodes are the same.
Phone screen · Coding · last seen 9 Apr 2026
Practise this problemMatches Lowest Common Ancestor of a Binary Tree. Open it in the code editor and run it against test cases.
Find the first 1 in a sorted binary matrix.
Onsite · Coding · last seen 9 Apr 2026
Compute how much rainwater can be trapped given an elevation map.
Onsite · Coding · last seen 9 Apr 2026
Practise this problemMatches Trapping Rain Water. Open it in the code editor and run it against test cases.
Design a live comment system for YouTube streams with millions of concurrent viewers, real-time delivery, loose ordering, spam filtering, and creator-pinned comments.
System design · System design · last seen 9 Apr 2026
Determine whether a string can become a palindrome after deleting at most one character.
Onsite · Coding · last seen 9 Apr 2026
Practise this problemMatches Valid Palindrome II. Open it in the code editor and run it against test cases.
Check whether a given abbreviation validly matches a word.
Onsite · Coding · last seen 9 Apr 2026
Describe a project you are most proud of.
Behavioral · Behavioral · last seen 9 Apr 2026
Describe a time something went wrong because of you.
Behavioral · Behavioral · last seen 9 Apr 2026
What do you do outside work to improve yourself?
Behavioral · Behavioral · last seen 9 Apr 2026
Implement a function that computes a number raised to an integer power efficiently.
Phone screen · Coding · last seen 8 Apr 2026
Practise this problemMatches Pow(x, n). Open it in the code editor and run it against test cases.
Describe a time you had a major disagreement with a cross-functional partner.
Behavioral · Behavioral · last seen 5 Apr 2026
Follow-ups: Did you deliver the impact the Product Manager committed to? If the feature was split, who absorbed the delayed metrics? How did you quantitatively justify the risk of pushing back?
Behavioral · Behavioral · last seen 5 Apr 2026
Design an ad aggregator service, covering how to scale with the number of servers and estimating database capacity.
System design · System design · last seen 13 Mar 2026
Given two linked lists whose nodes hold strings, determine whether both lists represent the same combined value.
Technical interview · Coding · last seen 13 Mar 2026
Given a sorted array, return the number that appears at least n/4 times.
Technical interview · Coding · last seen 13 Mar 2026
Asked two behavioral questions about the candidate's past projects.
Phone screen · Behavioral · last seen 12 Mar 2026
Simplify a Unix-style file path to its canonical form; write code, dry run it, and explain time complexity.
Phone screen · Coding · last seen 12 Mar 2026
Practise this problemMatches Simplify Path. Open it in the code editor and run it against test cases.
Find the level of a binary tree with the maximum sum of node values; write code, dry run it, and explain time complexity.
Phone screen · Coding · last seen 12 Mar 2026
Practise this problemMatches Maximum Level Sum of a Binary Tree. Open it in the code editor and run it against test cases.
Remove adjacent duplicate characters from a string, but the pairs that cancel are letters next to each other in the alphabet, such as AB or BA.
Technical interview · Coding · last seen 9 Mar 2026
Find the largest island you can form by flipping a single zero to one in a binary grid.
Technical interview · Coding · last seen 9 Mar 2026
Given a maze solver that uses breadth-first search, fix the failing tests in an AI-assisted round.
Other round · Coding · last seen 9 Mar 2026
Recent Meta interview experiences
Newest first. Each summary is ours; the full story is in the original post.
Earlier in 2026
Software Engineer · Mid-level
Phone screen, Technical interview, System design, Other round, Behavioral, Onsite · United States · 20 May 2026 · Rejected
The candidate, with three years of experience at a European large tech firm, interviewed for an E4 software engineer role at Meta in the US and was rejected after the full loop. The pre-loop screen and coding rounds went well, the system design round went solidly, and the AI-assisted coding round was only partly completed after a long delay on a bug. The recruiter declined to share feedback and advised waiting a year.
- Implement an LRU cache. (Practise this problem)
- Breadth-first traversal of a binary tree. (Practise this problem)
- Use BFS on a grid to reach the bottom-right cell and return the path taken. (Practise this problem)
- Compute the sum of a nested list where each element is multiplied by its depth level.
- Design a Ticketmaster-style ticket purchase flow for a Taylor Swift concert with only 20k tickets, handling scalability and strong consistency.
- In an AI-assisted coding round, fix a grid-traversal program to reach the end with directional arrows and walls, passing as many of five tests as possible.
Read the original on LeetCode Discuss →The interviewer asks the 3 questions from this report.
Run this Meta interview with the AI interviewerBusiness Engineer · Mid-level
Phone screen, Technical interview, Behavioral, System design, Other round · London · 10 May 2026 · Got an offer
The candidate interviewed for a Business Engineer L4 role in London and received an offer, later being offered a bump to L5 which was declined. The process included a coding screen, two data-structure rounds with follow-ups, a hiring manager behavioral round, a system design round on crowdsourced street imagery, and a cross-functional behavioral round. Key lessons were to move quickly to solutions and to document reasoning visibly.
- Sum the numbers formed by root-to-leaf paths in a binary tree.
- Count palindromic substrings in a string, with requests to optimize space and reach the expand-around-center approach. (Practise this problem)
- Find buildings with an ocean view; follow-ups: ocean on both sides, and choosing one building to remove to maximize ocean-view buildings.
- Given M sorted arrays with N total elements, find the Kth smallest element overall.
- Compute the average of each sliding window of size k in an array in O(n) time and O(1) space, handling invalid inputs. (Practise this problem)
- Find a local minimum in an array in sub-linear time, adapting for strict comparison and duplicates. (Practise this problem)
- Design a crowdsourced 3D street view using smartphone video from cars, showing different renderings by season or time of day.
- Tell me about a time style behavioral questions with deep follow-ups.
Read the original on LeetCode Discuss →The interviewer asks the 5 questions from this report.
Run this Meta interview with the AI interviewerSoftware Engineer, Infrastructure · Mid-level
Behavioral, System design, Technical interview, Other round · USA · 16 Apr 2026 · Waiting to hear back
The candidate completed four rounds of an E4 infrastructure software engineering loop at Meta and is awaiting feedback. They describe a smooth behavioral round, a system design covering requirements, entities, APIs and high-level design, an AI-assisted coding round with four test files, and a coding round with two medium problems. They felt mostly confident but missed one dry run and are unsure about expectations in design and AI rounds.
- Fix and extend code to pass four test case files, using AI assistance for some parts and explaining the approach for one implementation.
- Solve a medium problem involving a binary tree using depth-first search, explaining brute force, optimal solution, complexity and a dry run.
- Solve a medium binary search problem, explaining brute force, optimal solution, complexity and a dry run.
- Design a system, defining functional and non-functional requirements, core entities, API endpoints and a high-level architecture.
Read the original on LeetCode Discuss →The interviewer asks the 4 questions from this report.
Run this Meta interview with the AI interviewerProduct Software Engineer · Senior
Online assessment, Phone screen, Onsite, Technical interview, System design, Behavioral · 9 Apr 2026 · Waiting to hear back
The candidate, with about seven years of experience, interviewed for a product-oriented E5 position at Meta. The process began with two online assessments and a phone screen of two coding problems, followed by a four-round onsite covering coding, system design, and behavioral topics. They passed the onsite and moved to team matching. The post closes with takeaways on communication, clarifying questions, pacing, and prep resources.
- Merge overlapping intervals in a list; follow-ups on handling a real-time stream of intervals and duplicates. (Practise this problem)
- Find the lowest common ancestor of two nodes in a binary tree, including the case where both nodes are the same. (Practise this problem)
- Find the first 1 in a sorted binary matrix.
- Compute how much rainwater can be trapped given an elevation map. (Practise this problem)
- Design a live comment system for YouTube streams with millions of concurrent viewers, real-time delivery, loose ordering, spam filtering, and creator-pinned comments.
- Determine whether a string can become a palindrome after deleting at most one character. (Practise this problem)
- Check whether a given abbreviation validly matches a word.
- Describe a project you are most proud of.
- Describe a time something went wrong because of you.
- What do you do outside work to improve yourself?
Read the original on LeetCode Discuss →The interviewer asks the 6 questions from this report.
Run this Meta interview with the AI interviewerSoftware Engineer, ML
Phone screen, Technical interview · 8 Apr 2026 · Rejected
The candidate completed a technical phone screen for a machine learning software engineering position at Meta in March 2026. They were given two coding problems, clarified requirements, described their approach and complexity, coded solutions, and tested them, finishing early. Despite feeling confident, a rejection arrived three days later, and the recruiter did not respond to a feedback request. The candidate speculates about a hiring freeze.
- Implement a function that computes a number raised to an integer power efficiently. (Practise this problem)
- Given a binary tree, return its node values grouped level by level from left to right. (Practise this problem)
Software Engineer · Senior
Technical interview, System design, Behavioral · 5 Apr 2026 · Rejected
The candidate interviewed for a senior-level engineering position at Meta and was rejected despite strong coding performance and an acceptable system design round. The behavioral round was the downfall: a conflict story about compromising with QA on a release timeline was probed hard by the interviewer about impact, accountability, and risk justification. The candidate felt the story was too easily resolved and lacked senior-level scope, and regrets minimal behavioral preparation.
- Describe a time you had a major disagreement with a cross-functional partner.
- Follow-ups: Did you deliver the impact the Product Manager committed to? If the feature was split, who absorbed the delayed metrics? How did you quantitatively justify the risk of pushing back?
Read the original on LeetCode Discuss →The interviewer asks the 2 questions from this report.
Run this Meta interview with the AI interviewerSoftware Engineer
System design, Technical interview · 13 Mar 2026
The candidate interviewed for a Meta data collection team and described a design round plus a Python-based data structures and algorithms round. The design discussion centered on an ads aggregation service, including scaling by server count and database capacity estimates. The coding portion covered comparing linked lists, implementing a cache, and a sorted array frequency problem. No outcome, level, or location was shared.
- Design an ad aggregator service, covering how to scale with the number of servers and estimating database capacity.
- Given two linked lists whose nodes hold strings, determine whether both lists represent the same combined value.
- Implement an LRU cache. (Practise this problem)
- Given a sorted array, return the number that appears at least n/4 times.
Read the original on LeetCode Discuss →The interviewer asks the 3 questions from this report.
Run this Meta interview with the AI interviewerSoftware Engineer
Phone screen, Technical interview, Behavioral · 12 Mar 2026 · Waiting to hear back
The candidate describes a technical phone screen for a Meta software engineering role in March 2026. The interviewer began with two behavioral questions about the candidate's projects, then moved to two coding problems that required writing code, walking through it by hand, and analyzing complexity. The interviewer strictly enforced time limits for every question. The candidate is still waiting to hear back about the result.
- Asked two behavioral questions about the candidate's past projects.
- Simplify a Unix-style file path to its canonical form; write code, dry run it, and explain time complexity. (Practise this problem)
- Find the level of a binary tree with the maximum sum of node values; write code, dry run it, and explain time complexity. (Practise this problem)
Read the original on LeetCode Discuss →The interviewer asks the question from this report.
Run this Meta interview with the AI interviewerAndroid Engineer · Mid-level
Technical interview, Other round, Behavioral, System design · 9 Mar 2026 · Rejected
The candidate interviewed for a mid-level Android engineering position at Meta and was ultimately turned down. The coding portion featured one string-reduction task resembling a known problem with a twist on which letters cancel, plus a grid problem identical to a popular island question. An AI-assisted session involved repairing tests for a breadth-first maze solver. The behavioral conversation was standard. The design round gave only a brief prompt about an automatic wallpaper app, requiring heavy clarification and self-defined requirements.
- Remove adjacent duplicate characters from a string, but the pairs that cancel are letters next to each other in the alphabet, such as AB or BA.
- Find the largest island you can form by flipping a single zero to one in a binary grid.
- Given a maze solver that uses breadth-first search, fix the failing tests in an AI-assisted round.
- Design an Android app that sets the phone wallpaper to an image from the user's Facebook account whenever the phone is turned on, with no UI mocks provided.
Read the original on LeetCode Discuss →The interviewer asks the 4 questions from this report.
Run this Meta interview with the AI interviewerSoftware Engineer · Senior
Onsite, Technical interview, System design, Behavioral · 5 Mar 2026 · Rejected
The candidate interviewed for a senior-level software engineering role at Meta in the US and was rejected. The recruiter waived the phone screen, so the process went directly to an onsite loop. That loop covered a standard coding round, an AI-assisted coding round, a system design round, and a behavioral round. The recruiter phoned three days afterward with the decision but offered no feedback.
- Find two lines that, together with the x-axis, form a container holding the greatest amount of water. (Practise this problem)
- Find all words from a given list that can be formed on a letter grid by moving between adjacent cells. (Practise this problem)
- In an AI-enabled coding round, solve a problem about maximizing the number of unique characters.
- Design a live leaderboard system like the one used for LeetCode contests.
Read the original on LeetCode Discuss →The interviewer asks the 2 questions from this report.
Run this Meta interview with the AI interviewerCandidate report
Phone screen, Behavioral, Technical interview, Onsite, System design · 4 Mar 2026
The candidate prepares for a Meta final loop consisting of two coding rounds, one behavioral round and two architecture rounds, and asks which data structures and algorithms topics to prioritize. They also describe a completed screening stage that began with brief behavioral questions about failure and success, followed by two coding problems: designing a data structure and printing a binary tree in a particular layout. The outcome of the screen was a pass.
- Describe your biggest failed project.
- Describe your most successful project.
- Solve a data structure design problem.
- Print a binary tree in a specified printable format.
Read the original on LeetCode Discuss →The interviewer asks the 4 questions from this report.
Run this Meta interview with the AI interviewerSoftware Engineer · Mid-level
Online assessment, Phone screen, Technical interview, Other round, System design, Behavioral, Onsite · Bangalore, India · 1 Mar 2026 · Rejected
The candidate, a full-stack engineer with about four and a half years of experience, interviewed for a mid-level role at Meta in India. The process spanned several months and included an online assessment, a screening, and a full loop with coding, AI-enabled coding, product architecture and behavioral rounds. The candidate was rejected, reportedly mainly because of a weak behavioral signal, and shares preparation resources and lessons learned.
- Design an in-memory tag dictionary database with extensions for TTL and time-based backup, in four stages over 90 minutes.
- Remove the minimum number of parentheses to make a string valid, with a follow-up on improving memory use via a two-pass approach. (Practise this problem)
- Find the maximum area of an island in a binary grid. (Practise this problem)
- Print the left and right view (boundary) of a binary tree.
- Maximum length of a concatenated string with unique characters, with unit tests and optimization follow-ups, tested against a file of 12k+ wordle words.
- Design a Facebook post feed, focusing on APIs, user experience, pagination, localization and supporting different content types.
- Describe dealing with conflict involving multiple stakeholders.
- Describe a time you handled difficult feedback.
- Tell of a time you navigated an impactful task with little or no direction.
Read the original on LeetCode Discuss →The interviewer asks the 7 questions from this report.
Run this Meta interview with the AI interviewerSoftware Engineer · Senior
Phone screen, Onsite, Technical interview, System design, Behavioral · 17 Feb 2026
A mid-career engineer with four years of experience describes a standard Meta E5 loop: one phone screen followed by four onsite sessions covering two coding rounds, system design, and behavioral. The candidate recounts the problems faced and per-round verdicts, noting a moment of near-freezing in a graph problem. The post also promotes an AI interview-prep tool, and the final result is not stated.
- Variation of the valid palindrome problem where one character may be deleted. (Practise this problem)
- Find the kth largest element in an array, with quickselect expected. (Practise this problem)
- Graph traversal problem on a grid with obstacles that can be eliminated, with an extra constraint added by the interviewer.
- Merge k sorted linked lists, with follow-up questions on heap memory usage. (Practise this problem)
- Design a notification system that can alert one million users during a flash sale, focusing on fan-out and queue handling.
- Tell me about a time you had a conflict.
Read the original on LeetCode Discuss →The interviewer asks the 3 questions from this report.
Run this Meta interview with the AI interviewerSoftware Engineer · Mid-level
Online assessment, Behavioral, Technical interview · India · 4 Feb 2026 · Waiting to hear back
The candidate took an online assessment with coding and behavioral parts for an E4 software engineer position at Meta in India. They completed three full levels of the coding portion and passed 21 of 34 test cases, then asked how likely that was to advance. In an update, they reported receiving an interview invitation, with interviews scheduled for the following week.
Read the original on LeetCode Discuss →Software Engineer · Senior
Technical interview · 2 Feb 2026 · Waiting to hear back
The candidate interviewed for an E5 role at Meta and went through a coding round after a short introduction. The first problem involved finding the shortest substring meeting given constraints, which they solved by moving from brute force to a sliding window with frequency tracking, and they discussed edge cases. The second was a known random index picking problem. They felt positive and were awaiting the outcome.
- Find the shortest substring that satisfies given input constraints, a variation of a substring search problem; discuss edge cases such as empty input or no valid result.
- Randomly pick the index of a target value from an array that may contain duplicates.
Read the original on LeetCode Discuss →The interviewer asks the 2 questions from this report.
Run this Meta interview with the AI interviewerSoftware Engineer · Senior
Phone screen · London · 12 Jan 2026 · Waiting to hear back
The candidate describes a phone screen for an IC4 position in London at Meta. The session had two coding problems: one involving computing how long each function actively ran from begin and end timestamp logs, and another on searching a binary search tree for a value nearest a target. Feedback was positive, and a full interview loop was to be scheduled.
- Given log entries with function name, timestamp, and begin or end marker, compute each function's exclusive running time, excluding time spent in nested calls to other functions.
- Given the root of a binary search tree and an integer target, return the node whose value is closest to the target.
Read the original on LeetCode Discuss →The interviewer asks the 2 questions from this report.
Run this Meta interview with the AI interviewer
2025
Candidate report
Online assessment · 19 Dec 2025
The candidate describes a Meta online assessment taken on CodeSignal, structured as four progressive levels building a task management system. Early levels cover creating, updating and retrieving tasks, then searching and listing them with sorting rules, followed by user quotas and task assignment over time. The final level involved completing tasks plus another function the candidate could not recall. No outcome is mentioned.
- Level 1: Implement a task manager with addTask returning auto-incremented IDs, updateTask returning success boolean, and getTask returning name and priority as JSON.
- Level 2: Implement searchTask by name filter sorted by priority descending then task ID ascending with a limit, and taskList sorted by ID then timestamp with a limit.
- Level 3: Add users with a quota of active tasks, assign tasks to users when quota is available and freed after finish time, and return a user's tasks active at a given timestamp.
- Level 4: Implement completeTask to mark a task as completed, plus one additional unremembered function.
Read the original on LeetCode Discuss →The interviewer asks the 4 questions from this report.
Run this Meta interview with the AI interviewerSoftware Engineer · Senior
Onsite, Technical interview, Other round, System design, Behavioral · 17 Dec 2025
The candidate described an E5 infrastructure onsite at Meta with four parts. The coding round had two standard problems and a friendly interviewer who asked for code explanation and testing approach. The AI-assisted round was unpleasant: the work centered on passing unit tests, and the candidate could not finish the solver. A system design round on Instagram and a behavioral round followed. The candidate felt the AI round hurt their chances.
- Merge three sorted arrays into one sorted array.
- Return the right side view of a binary tree. (Practise this problem)
- Given a list of words, find a subset that uses the most characters without repeating any letter across the chosen words; solved against unit tests.
- Design Instagram covering posts, images and friends; discuss consistent hashing and handling high read volume in the feed service, including fan-out.
- Behavioral topics: conflict with a coworker, negative feedback from a manager, prioritization, and ambiguous requirements.
Read the original on LeetCode Discuss →The interviewer asks the 4 questions from this report.
Run this Meta interview with the AI interviewerMachine Learning Engineer · Staff
Technical interview, System design · United States · 14 Dec 2025 · Waiting to hear back
The candidate interviewed for a senior machine learning engineering position at Meta in the US, at the IC6 level. The coding sessions reportedly used familiar Meta-tagged practice problems without changes and were finished quickly. The ML design portion covered a general embedding platform and an Instagram video recommender. No result had arrived, but the candidate had already accepted a different offer elsewhere.
- Design an ML platform that produces general-purpose content embeddings used by downstream tasks such as classification, recommendation, and search.
- Design a video recommendation system for Instagram.
Read the original on LeetCode Discuss →The interviewer asks the 2 questions from this report.
Run this Meta interview with the AI interviewerSoftware Engineer · Mid-level
Online assessment, Phone screen, Onsite, Technical interview, System design, Behavioral · 14 Dec 2025 · Got an offer
The candidate describes a six-stage hiring loop for a mid-level engineering position in the United States. It began with a CodeSignal assessment involving an incremental file system task, followed by a phone screen with two quick problems on flight pricing and tree ordering. The virtual onsite included a coding session with string permutations and a tree-average variant, an AI-flavored coding exercise, a food delivery platform architecture discussion, and a values-focused conversation. The author reports receiving an offer and plans to start in early 2026.
- Progressive file system implementation problem with multiple levels; the candidate solved three of four levels.
- Find the cheapest round-trip cost for a given day using an array of prices.
- Perform a vertical order traversal of a binary tree.
- Generate all letter case permutations of a string.
- Count nodes whose value equals the average of their subtree, plus a variation of it.
- Design DoorDash, covering order lifecycle, scalability, availability, real-time updates, and trade-offs.
Read the original on LeetCode Discuss →The interviewer asks the 6 questions from this report.
Run this Meta interview with the AI interviewerMachine Learning Engineer · Mid-level
System design, Technical interview · 3 Dec 2025 · Waiting to hear back
The candidate describes an ML system design round for a machine learning engineer position at E4 level. The prompt asked for designing a recommendation system for Instagram. Time ran short, so evaluation approaches and model improvements were covered only briefly without the depth desired. The candidate felt coding and AI-enabled coding rounds went reasonably well and asks the community whether an offer is still plausible despite an imperfect design round.
- Design a recommendation system for Instagram, including how to evaluate it and improve the models.
Read the original on LeetCode Discuss →The interviewer asks the question from this report.
Run this Meta interview with the AI interviewerSoftware Engineer (SDE2) · Mid-level
Phone screen, Onsite, Technical interview, System design, Behavioral, Other round · 26 Nov 2025 · Rejected
The candidate went through a full Meta loop for a mid-level engineering role and was ultimately turned down without receiving feedback. A phone screen with two coding problems went smoothly and led to the onsite. The onsite had a shaky coding round with an interrupting interviewer, a new AI-assisted debugging and string task, a social feed architecture discussion with unexpected UI questions, and a behavioral session that began nervously but improved.
- Add two numbers represented as strings and return the sum as a string. (Practise this problem)
- Return the k most frequently occurring elements from an array. (Practise this problem)
- Given a sorted array, return the squares of each number in sorted order. (Practise this problem)
- Evaluate a string arithmetic expression containing integers and operators without parentheses. (Practise this problem)
- In an AI-enabled round, debug existing code, then implement a function that removes words from a list when they share letters with others.
- Design a newsfeed or Instagram-like system with posts, likes and comments, plus some unexpected UI-related follow-ups.
- Standard behavioral questions, about five or six, with follow-ups on a few.
Read the original on LeetCode Discuss →The interviewer asks the 3 questions from this report.
Run this Meta interview with the AI interviewerML Software Engineer · Senior
Technical interview, System design, Behavioral · 26 Nov 2025 · Got an offer
The candidate interviewed for an E6 machine learning software engineering position at Meta. The process included two coding rounds with array and heap style problems plus a custom token-issuing design task, a system design round on video feed and ranking, and a behavioral round. The result was a downlevel offer to E5, which the candidate declined because of other opportunities in tech and finance.
- Find the k points closest to the origin. (Practise this problem)
- Two-sum style variant: determine whether a subarray summing to K exists and return the length of the longest such subarray. (Practise this problem)
- Find the k-th largest element in an array, with a variant returning its index instead of the value. (Practise this problem)
- Issue tokens from a 16-bit integer space; when tokens are returned, the smallest returned token must be issued next. Handle edge cases. (Practise this problem)
- Design an Instagram video feed end to end.
- Design feed ranking for Facebook, focusing on the second-stage ranker.
- Behavioral topics such as conflict resolution and pivoting.
Read the original on LeetCode Discuss →The interviewer asks the 3 questions from this report.
Run this Meta interview with the AI interviewerProduct Software Engineer · Mid-level
Online assessment, Phone screen, Onsite, Technical interview, System design, Behavioral · 16 Nov 2025 · Waiting to hear back
The candidate describes a Meta E4 process starting with a four-stage coding assessment and a work-style questionnaire, followed by a phone screen and four onsite rounds covering coding, a system design, and behavioral topics. Feedback stressed communication and edge cases. The candidate cleared the interviews and moved to team matching, and shares takeaways about clarity, asking questions, accepting hints, and managing time.
- Count the ones in a subarray of a large binary array given start and end indices; follow-ups on digits 0-9 and start greater than end.
- Find the lowest common ancestor of two nodes in a binary tree, including edge cases where both references are the same node. (Practise this problem)
- Find the first 1 in a sorted binary matrix.
- Compute trapped rain water given elevation heights. (Practise this problem)
- Design a small-scale concert ticket booking service with one venue and many concerts, covering requirements, API, data model and bottlenecks.
- Determine whether a string can be a palindrome after deleting at most one character. (Practise this problem)
- Check whether a word matches a given abbreviation.
- Tell about a project you are most proud of and your role in it.
- Describe a time you pushed a bug to production or caused something to go wrong, and what you learned.
- What do you do outside work to learn or improve yourself?
Read the original on LeetCode Discuss →The interviewer asks the 7 questions from this report.
Run this Meta interview with the AI interviewerData Engineer, Product Analytics · Mid-level
Phone screen · 11 Nov 2025 · Waiting to hear back
The candidate, a data engineer with roughly four years of experience, describes a short initial recruiter call for a product analytics data engineering position. It covered background fit, current responsibilities, and three rapid technical questions, one on SQL and two on Python. Poor audio caused some stumbles on the Python items, though the recruiter explained the approach afterward. The candidate asks how much weight this call carries and how long next steps take.
- A quick SQL question asked during the recruiter call (specific content not given).
- Two quick Python questions asked during the recruiter call (specific content not given).
- Why are you interested in this role?
Read the original on LeetCode Discuss →The interviewer asks the 3 questions from this report.
Run this Meta interview with the AI interviewerData Engineer, Product Analytics · Mid-level
Phone screen · 11 Nov 2025 · Waiting to hear back
The candidate describes a brief recruiter call with Meta for a product analytics data engineering position. The roughly eleven minute conversation covered background fit, current responsibilities, motivation for the role, and three quick technical questions, one on SQL and two on Python. The candidate handled the SQL item well but struggled with the Python ones because of poor audio. The candidate asks how much weight this stage carries and how long next steps take.
- A quick SQL question asked during the recruiter call (specific content not given).
- Two quick Python technical questions asked during the recruiter call (specific content not given).
- Why are you interested in this role?
Read the original on LeetCode Discuss →The interviewer asks the 3 questions from this report.
Run this Meta interview with the AI interviewerSoftware Engineer, Product
Onsite, Technical interview, System design, Behavioral · Menlo Park · 8 Nov 2025
The candidate describes an onsite loop at Meta that included an AI-assisted coding round, a conventional coding round, system design, and behavioral. In the AI round, they worked in a multi-file codebase, fixed failing unit tests, then implemented a stubbed function using a lightweight model. The interviewer probed complexity, revealing the solution was exponential because the task was NP-complete. The candidate found the format practical and the interviewer friendly.
- Fix two failing unit tests among four in a utility module of a small multi-file codebase, caused by minor logic errors, using debugging or AI assist.
- Implement a stubbed function that takes a list of words and finds a subset meeting a given property; then discuss time and space complexity and possible optimization (the problem was NP-complete).
Read the original on LeetCode Discuss →The interviewer asks the 2 questions from this report.
Run this Meta interview with the AI interviewerSWE Product
Onsite, Technical interview · USA · 8 Nov 2025
The candidate describes a non-AI coding session during a Meta onsite for a product software engineering role. Two problems came up: a standard binary tree vertical ordering task, and an ambiguous friend-request counting puzzle based on age rules that required clarifying requirements. The candidate did not finish a fully working solution for the second problem but explained their reasoning, and later found it manageable once framed properly.
- Given a binary tree that is not a search tree, output node values column by column from leftmost to rightmost, with higher nodes listed before lower nodes within the same column.
- Given a list of centaur ages, count valid directional friend requests under age-based rules: under-100s can befriend under-100s, over-100s can befriend strictly older over-100s, or those older than half their age plus seven. Any satisfied rule makes a request valid.
Read the original on LeetCode Discuss →The interviewer asks the 2 questions from this report.
Run this Meta interview with the AI interviewer
Wrote one of these posts and want it taken down? Tell us and we'll remove it.
Questions, answered
- How many rounds are in the Meta interview?
- It depends on the level: E4 (Software Engineer) has 4 rounds; E5 (Senior Software Engineer) has 4 rounds. Each level page lists every round with its length and format.
- What questions does Meta ask in interviews?
- The ones candidates reported most recently: LRU Cache; Binary Tree Level Order Traversal; Shortest Path in Binary Matrix. The full list is on this page and grows with each refresh.
- Can you use AI in the Meta interview?
- Many Meta technical interviews now include an AI assistant built into the CoderPad environment, and candidates are expected to use it. No outside AI tools or assistance are authorized.
- How often is this page updated?
- Every 7 days. Each refresh looks for new public Meta interview posts, checks the link works and adds what it finds. Old reports stay. The last refresh finished on 9 Nov 2026; the next is due on 16 Nov 2026.
- How do I practise these Meta questions?
- A question with a practice button matches a problem in our catalogue. The button opens the AI interviewer on that problem with Meta set as the company: it asks the question, follows up on your approach and grades you at the end.
- Is this the official Meta interview process?
- No. Thita is not affiliated with Meta. The loops by level are checked against current public descriptions of the process; the reports are candidates' own accounts, so treat them as what people saw recently.
Get new Meta interview questions every week
When a refresh finds new Meta reports, you get the questions and a link to each post. Nothing else.