Meta SWE Intern OA 2027: Coding Questions, Preferences at Work, and What Comes Next

Updated August 16, 2026: A Meta online assessment invitation is exciting, but it also creates three immediate questions: What will the coding section look like, does Preferences at Work affect the decision, and what happens if you pass?

Here is the evidence-based answer. Recent Meta intern cycles commonly included a timed coding assessment and a separate workplace-preferences questionnaire, followed by recruiting review and a live technical interview for selected candidates. Meta has not published one universal Summer 2027 OA blueprint, so the platform, timer, question count, deadline, and allowed resources in your own invitation take priority over every forum post.

Prepare for two different signals: timed coding execution and consistent workplace preferences.

Meta SWE Intern OA 2027: Quick Verdict

Prepare for a fast coding screen, but do not stop at coding. Candidate reports from recent intern cycles describe a multi-question online coding assessment plus Preferences at Work. Candidates who advanced then reported a live coding interview, often with two problems and a strong emphasis on explaining decisions.

The safest preparation strategy has three parts: practice common Meta coding patterns under a timer, answer workplace-preference questions honestly and consistently, and continue preparing for a live technical screen immediately after submission. A perfect-looking OA is not a public guarantee of an interview because recruiting can also consider resume fit, location, role requirements, and available headcount.

What Is Confirmed, Reported, and Still Unknown?

Meta's public software engineering technical screen guide is useful for understanding the live interview. It says interviewers assess communication, problem solving, coding, and verification, and that candidates should expect roughly two problems in about 40 minutes. That is official guidance for a technical screen, not proof of the 2027 OA configuration.

Recent candidate reports provide the missing assessment context. A reported Meta intern process included a coding assessment, Preferences at Work, and then an online technical interview. Other 2025 reports mention a four-question OA followed by a 45-minute live coding round. These reports show a recurring pattern, but they cannot establish a universal cutoff or identical format across countries and recruiting batches.

Detail	Evidence level	How to plan
Timed coding assessment	Commonly reported in recent Meta intern cycles	Practice algorithms in a browser editor with a strict timer.
Preferences at Work	Reported by multiple intern candidates	Treat it as a separate signal and answer consistently.
Four coding questions	Reported in recent cycles, not confirmed for every 2027 invite	Use it as one simulation format, not a promise.
Platform and duration	Can vary by invitation and location	Use the exact instructions in your email and setup page.
Passing score	No public universal Meta cutoff	Maximize correctness and testing; ignore rumor-based guarantees.

What Does a Meta OA Invitation Actually Mean?

An invitation means Meta wants another signal about your candidacy. It is better than silence, but it does not necessarily mean your resume has completed every later review. Some companies send assessments broadly, while others apply more screening before the invite; Meta does not publish a rule that lets candidates infer exactly where they stand from the email alone.

Read the deadline carefully and open the assessment's system-check page early. Confirm the coding language options, whether navigation between questions is allowed, what resources are permitted, and whether any identity or integrity checks apply. Do not import rules from an unrelated CodeSignal or HackerRank assessment just because the interface looks familiar.

The coding section, Preferences at Work, and recruiting review answer different questions about a candidate.

What Coding Questions Should You Expect?

Meta has a reputation for speed-oriented algorithm interviews. The useful preparation target is not an obscure trick; it is recognizing a familiar data shape quickly, selecting a suitable data structure, writing correct code, and checking the edges before the clock expires.

Build your base around arrays, strings, hash maps, sets, sorting, two pointers, sliding windows, stacks, queues, trees, graphs, and backtracking. For each solution, be able to state the expected time and space complexity without turning the explanation into a lecture. A concise plan followed by clean implementation is more valuable than naming five patterns and finishing none.

How difficult is the Meta intern OA?

Candidate descriptions vary from straightforward to time-pressured. That variation may reflect different question sets, candidate strengths, or recruiting regions. A realistic benchmark is being able to solve common easy and medium problems reliably, while making useful progress on a harder final problem without destroying working code.

Do not measure readiness by whether you recognize the prompt. Measure whether you can derive the solution, implement it from a blank editor, test it, and explain why it is correct. That skill transfers even when Meta changes the wording or combines two familiar ideas.

Do visible test cases guarantee a pass?

No. Visible tests confirm only the examples you can see. Hidden cases may include empty input, duplicates, unusual ordering, large values, repeated states, disconnected graphs, or performance limits. Passing every sample can still leave an off-by-one error or an algorithm that times out at scale.

Reserve the final minutes for verification. Re-read the constraints, test the smallest valid input, test duplicated values, check output ordering, and compare the implementation's complexity with the largest input. Verification is also one of the four areas named in Meta's official live technical-screen guide.

How to Approach Preferences at Work

Preferences at Work is not another coding section. Candidate reports describe it as a behavioral-style questionnaire about the environments, trade-offs, and ways of working you prefer. The exact number and wording can change, so memorizing an unofficial answer key is fragile and can create inconsistent responses.

Answer as the strongest truthful version of yourself. Read each choice literally, distinguish what you usually do from what sounds admirable, and keep your responses consistent across similar scenarios. If you value both independence and collaboration, use the wording of the specific situation rather than selecting one identity for the entire test.

A useful review is Meta's public values and the behaviors your own projects can prove: ownership, moving with appropriate speed, direct communication, learning from mistakes, and building for impact. Do not manufacture a persona. Later interviewers can ask for concrete examples, and contradictions are harder to defend than a nuanced, honest preference.

A Five-Day Meta OA Preparation Plan

This plan assumes you already know one interview language. If your deadline is shorter, combine Days 1 and 2, but keep the full simulation and final review.

Day	Focus	Concrete output
1	Meta diagnostic	Record time, bugs, and missed edge cases.
2	Core speed	Practice arrays, strings, hashing, sorting, and two pointers in short timed sets.
3	Deeper patterns	Complete trees, graphs, and backtracking problems; rewrite one weak solution from memory.
4	Full simulation	Use the exact timer from your invitation. Work without external help and reserve time for testing.
5	Verification and recovery	Fix the simulation, review your language template, check the environment, and sleep normally.

Diagnose with real questions first, then build speed, depth, and verification.

How to Execute During the Coding Assessment

Scan the full set before committing

If the platform allows navigation, read every prompt and constraint first. Start with the problem whose solution path is clearest, not necessarily the first one. Early working code reduces pressure and gives you more time to reason about the hardest task.

Write the invariant before the code

State what must remain true: the window contains no duplicate, the queue represents the next BFS layer, or the prefix count identifies uniqueness. A stable invariant keeps you oriented when a sample passes but a hidden case fails.

Protect working solutions

Do not launch a large rewrite in the final minutes unless the current complexity cannot pass. Improve one risk at a time, rerun the tests, and preserve a known-good version when the editor supports it. Partial progress on a difficult question is usually better than breaking two completed ones.

What Happens After You Submit?

Your result may enter recruiting review rather than producing an immediate pass or fail screen. Response times can vary by team, location, recruiting volume, and the order in which applications are reviewed. A short wait is not evidence of rejection, and a perfect visible-test result is not evidence of an automatic interview.

Candidates who advance should be ready for a live technical conversation. Meta's official guide says the coding screen evaluates communication, problem solving, coding, and verification, with roughly two problems in about 40 minutes. Unlike an OA, the interviewer can interrupt, ask for a better approach, or probe a missed edge case.

Keep one timed coding session in your routine while you wait. Also prepare concise stories about collaboration, conflict, mistakes, ownership, and impact using behavioral and leadership questions.

Common Mistakes That Lower Otherwise Strong OA Performance

The first mistake is practicing only recognition. Reading solutions feels efficient, but the OA requires recall, implementation, and debugging from a blank editor. Attempt first and compare second.

The second mistake is optimizing before confirming correctness. A simple solution that fits the constraints and passes edge cases is stronger than an elegant idea with unfinished code. Read the limits before deciding how much optimization is necessary.

The third mistake is treating Preferences at Work as a puzzle with one hidden personality. Over-engineered answers can become inconsistent. Read carefully, answer honestly, and save your strategic energy for the coding and later interviews.

Meta SWE Intern OA 2027 FAQ

Does every Meta SWE intern applicant receive an OA?

Meta does not publish a rule saying every applicant receives the assessment. An invitation is a positive process signal, but it does not reveal whether every resume review is complete or how broadly that recruiting batch sends OAs.

How many coding questions are on the Meta intern OA?

Recent candidates have reported four-question assessments, but Meta has not published one universal 2027 count. Use four questions as a possible simulation format and follow the exact count and timer shown in your invitation.

What score is needed to pass?

No public universal cutoff is available. Aim for correct, efficient, well-tested solutions, but remember that recruiting may also review role fit and available capacity. Do not treat a rumored score as an interview guarantee.

Is Preferences at Work important?

It should be treated as a real assessment signal, not a survey to click through. Answer honestly and consistently, read every comparison carefully, and avoid copying an unofficial answer pattern that does not reflect how you actually work.

How long does Meta take to respond after the OA?

There is no dependable universal timeline. Candidates can hear back on different schedules based on region, team, and recruiting volume. If the timeframe in your email has passed, send one concise follow-up; otherwise continue preparing for the technical screen.

Can I use AI, notes, or another browser tab?

Only use resources explicitly allowed by your assessment instructions. Meta's later AI-enabled coding format should not be interpreted as permission to use an AI assistant during a traditional OA. When a rule is unclear, ask the recruiting contact before the test.

Prepare for the Assessment You Actually Received

The strongest Meta OA strategy is deliberately unglamorous: read your rules, recognize common patterns, write correct code, test the boundaries, and answer workplace-preference questions consistently. Rumors can help you form a simulation, but your invitation controls the real test.

If the OA goes well, you will already be preparing for the communication, problem-solving, coding, and verification signals Meta evaluates next.

Research note: Process details were checked on August 16, 2026 against Meta's public technical-screen materials and recent candidate reports. Assessment configurations can change, so the instructions in your invitation take priority.