Meta ML Engineer candidates report a coding screen followed by a loop of coding, ML system design, and behavioral rounds. Some also mention product sense and team matching. Prepare for fast LeetCode-style coding plus ranking and integrity design.
$227K
Avg. Base Comp
$473K
Avg. Total Comp
5-6 rounds
Typical Rounds
Not reported
Process Length
Candidates describe the Meta ML Engineer interview as heavier on software-engineering coding than the title suggests. One candidate saw two LeetCode-style problems in 45 minutes in the screen, covering a binary search tree, DFS, and string or parentheses questions. Another had a 40-minute phone screen with one easy and one hard problem, and a third was given a longest-subarray variant with a fixed-memory streaming follow-up. Getting to a correct solution quickly is the recurring theme, and one candidate noted that one round had no terminal to run code.
The loop reports differ in size. One candidate describes four onsite rounds: two coding, one ML system design, and one behavioral. Another was told to expect five rounds over two days: ML design, two coding rounds, product sense and metrics, and behavioral. A third reports four loop rounds after the screen, including an unfamiliar AI-assisted coding round where the task was to find and fix a bug, then implement and improve a solution.
The ML design prompts are tied to Meta-style products. Candidates mention recommendation and ranking, feed freshness with a cold-start follow-up for new creators, and harmful content detection. The Reels engagement question asked which specific features to pull first, so prepare concrete metrics and a clear order of investigation.
One post-onsite report says the next step after passing was team matching, which could take a while because open roles were limited. Candidate outcomes in this set were mostly rejections or withdrawals, so treat the process as demanding. Total timelines are not reported.
Synthesized from 5 candidate reports by our editorial team.
Had an interview recently?
Share your experience. Unlock the full guide.
Real interview reports from people who went through the Meta process.
I completed the ML Engineer onsite at Meta on April 3 and was still waiting to hear back afterward. The recruiter had explained that, if I passed, the next stage would be team matching rather than an immediate placement, and that matching could take a while because there were relatively limited open roles. At the time of posting, I had not received any outreach from hiring managers or team-match requests. The recruiter had mentioned about two weeks earlier that they were wrapping up interviews and expected to make a decision soon, so the silence was frustrating but not necessarily a rejection. My main takeaway is to stay in touch with your recruiter after the onsite and ask directly where you are in the decision or team-matching process; a delay alone does not appear to mean a no.
Prep tip from this candidate
After the onsite, ask your recruiter whether you are awaiting an interview decision or have moved into team matching, since open-role availability can extend the timeline.
Share your own interview experience to unlock all reports, or subscribe for full access.
Sourced from candidate reports and verified by our team.
Topics based on recent interview experiences.
Featured question at Meta
Write a query to return whether each user's subscription date range overlaps with any other completed subscription
| Question | |
|---|---|
| Merge Sorted Lists | |
| Weighted Keys | |
| Find the Missing Number | |
| Scrambled Tickets | |
| 500 Cards | |
| Find Bigrams | |
| Search Ranking | |
| One Element Removed | |
| P-value to a Layman | |
| The Brackets Problem | |
| Amateur Performance | |
| Detecting Firearm Sales | |
| Get Top N Frequent Words | |
| Level Of Rain Water In 2D Terrain | |
| Nearest Common Ancestor | |
| Recurring Character | |
| Facebook Stories | |
| Impression Reach | |
| Bank Fraud Model | |
| Random Forest Explanation | |
| Lazy Raters | |
| Type-ahead Search | |
| Find the First Non-Repeating Character in a String | |
| Binary Tree Conversion | |
| Precision and Recall | |
| Flatten JSON | |
| Fill None Values | |
| Booking Regression | |
| Find Duplicate Numbers in a List |
Synthesized from candidate reports. Individual experiences may vary.
Candidates report a recruiter call outlining the loop, then a technical screen that moved straight into coding. Reports include two LeetCode-style problems in 45 minutes and one easy plus one hard problem in 40 minutes. Topics mentioned include trees, DFS, strings, and sliding-window or streaming variants.
Candidates report two coding rounds in the loop, with several short problems rather than one long one. Topics mentioned include graphs, dynamic programming, trees, and word-ladder-style shortest paths with follow-ups. One candidate had no terminal for running code, so practice writing bug-free code unaided.
One candidate reports an AI-assisted coding round in an unfamiliar format. The task was to find and fix a bug, then implement a solution and improve it. Other candidates did not mention this round, so it may not appear in every loop. Practice debugging unfamiliar code and explaining your improvements aloud.
Candidates describe ML design prompts around recommendation and ranking, feed freshness, and harmful content detection. Follow-ups probed cold-start handling, evaluating a fallback without contaminating the main ranking metrics, and choosing specific features. Prepare to discuss data, modeling, evaluation, and tradeoffs for Meta-scale products.
One candidate reports a product sense and metrics round, including a diagnosis question about a drop in Reels engagement. Behavioral rounds were reported by several candidates, one describing it as conversational. A behavioral question about pushing back on a technical decision asked what the candidate would do differently now.
One candidate says the recruiter explained that passing the loop leads to team matching, which could take a while because open roles were limited. That candidate had heard nothing from hiring managers after the onsite. Ask your recruiter whether you are awaiting a decision or in matching.
