# Meta Machine Learning Engineer (MLE) Interview Guide

Aced (formerly Exponent) publishes a public search API over its indexed pages; the guide is at https://www.aced.io/agents/agent-access.

Source: https://www.aced.io/guides/meta-machine-learning-engineer-interview
Source observed: 2026-11-04T21:21:27.052Z
Last checked: 2026-11-04T21:21:27.052+00:00
Indexed: 2026-11-05T03:33:11.835+00:00

---

# Meta Machine Learning Engineer (MLE) Interview Guide

Updated 5 days ago by Meta candidates. We provide this expert guide through Aced, formerly Exponent, to help you understand and prepare for Meta’s machine learning engineer interview. It covers generalist MLE roles and specialized tracks such as computer vision, including process stages, interview formats, evaluation criteria, reported questions, preparation guidance, and related Aced resources.

**Guide sections:** [Expert guide](https://www.aced.io/guides/meta-machine-learning-engineer-interview) · [Interview experiences (13)](https://www.aced.io/guides/meta-machine-learning-engineer-interview/experiences) · [Questions (121)](https://www.aced.io/guides/meta-machine-learning-engineer-interview/questions) · [Resumes](https://www.aced.io/guides/meta-machine-learning-engineer-interview/resumes)

> **Verified**
> 
> We create our guides from recent, real, first-hand insights shared by interviewers and candidates. If your experience differs, [tell us here](https://www.aced.io/questions/contribute).

Meta’s MLE interview varies by specialization. Generalist candidates and specialized candidates receive different coding and design rounds. “AI coding” can mean an LLM-assisted debugging session in an unfamiliar codebase or ML math implementations such as sparse matrix-vector multiplication, depending on the track. Every format tests the ability to write correct code under time pressure, making timed coding practice central to preparation.

## Meta MLE interview process

The process includes a recruiter screen or written pre-screen, a technical screen, a final loop, and team matching. Meta lists generalist MLE roles under **Software Engineer, Machine Learning**; specialized tracks use a modified process.

1. **Recruiter screen or pre-screen:** Written questionnaire on Meta’s candidate portal, recruiter call, or hiring-manager screen on specialized tracks.
2. **Technical screen:** A 45-minute live coding interview with 2 challenges, or a 60-minute AI-native screen for some senior candidates.
3. **Final loop:** 4–5 virtual interviews covering coding, design, and behavioral topics, sometimes across several days.
4. **Team matching:** Placement on a specific team after passing the final loop.

The process generally takes roughly **4–8 weeks** from first contact to a hiring decision; team matching can take longer. Meta builds an AI assistant into many coding and design interviews. Practice in the CoderPad environment available through your Meta Career Profile before the screen.

### Generalist and specialized final loops

| Loop element | Generalist MLE (Software Engineer, Machine Learning) | Specialized track (for example, computer vision) |
|---|---|---|
| Screening | Pre-screen questionnaire or recruiter call, then a technical screen | Recruiter resume review, then a hiring-manager screen; some loops add a technical screen |
| Coding rounds | One standard coding round plus one AI-enabled coding round; some loops use 2 standard or 2 AI-enabled rounds | One standard coding round plus one ML coding round |
| Design rounds | One ML system design round; E6 and some E5 loops include 2 | One AI system design round plus one system design round |
| Behavioral | One behavioral round | One behavioral round |
| Total interviews | 4–5 | 5 |

The process is team-independent until team matching, but your track determines the coding and design rounds. Ask your recruiter which rounds are scheduled and align your preparation accordingly.

## Recruiter screen and pre-screen questionnaire

The pre-screen checks whether your experience matches the target level. Depending on the loop, it may be a written questionnaire, recruiter call, or hiring-manager screen.

The portal questionnaire is mostly open-ended questions about your current team, tools, and role, with some multiple-choice questions. Answer in full paragraphs and include team size, scope, and tools. For many E5-and-above candidates, it focuses on day-to-day AI-tool use, keeping current with tooling, and judging AI output quality. Include specific tools, workflows, and examples of AI mistakes you caught.

Recruiter calls verify relevant experience against the target level. Specialized-track hiring managers may ask about each past project, what you delivered, and its impact; prepare a summary of each recent project ending with a measurable result.

Interviewers look for:

- **Level fit:** Experience and scope appropriate to the target level.
- **Project impact:** What you delivered and how you measured the result.
- **Technical depth:** Motivation, assumptions, trade-offs, and limitations behind your work.
- **Tools and AI use:** Tools, programming languages, and AI tools used in your current role.
- **Role mix:** How you divide time between hands-on coding and leadership.

### Recently asked questions

- What is the size of your current team, and what does it work on?
- Which tools, programming languages, and AI tools do you use day to day?
- What percentage of your current role is coding versus leadership?
- Walk through each recent project. What did you deliver, and what was the impact?
- How many years of experience qualify you for this level?
- How do you use AI tools day to day, and how do you judge the quality of their output?

Practice project walkthroughs and coding answers with our [Meta interview questions](https://www.aced.io/questions?company=meta-facebook). Meta interviewers mostly use shared question banks, so prompts can recur across candidates and rounds.

## Technical screen

The standard technical screen is a **45-minute** live CoderPad interview with **2 data-structures-and-algorithms challenges**. You are expected to explain your reasoning and generally need complete, bug-free solutions to advance. Code execution is off in the standard screen, so trace the first solution line by line before beginning the second. Reserve time for edge cases and runtime and space complexity. Senior candidates may also receive behavioral questions; prepare 2 stories.

Some E5-and-above candidates take a **60-minute AI-native screen** with an AI assistant instead of the standard format. Ask your recruiter which version applies.

Interviewers look for presentation clarity, code traceability, complexity analysis, code quality, and correctness.

### Recently asked questions

- Validate a Unix-style file path containing slashes and dots.
- Merge 2 sorted lists of intervals.
- Return every character in string A that also appears in string B.
- Find the shortest path between 2 points in a grid.
- [Merge 2 sorted linked lists into a single sorted linked list](https://www.aced.io/questions/5566/merge-two-sorted-linked-list).
- [Find the k closest points to the origin](https://www.aced.io/questions/2095/find-points-closest-origin).
- Count occurrences of a target integer in a sorted array in **O(log N)** time.

## Coding interview

The final-loop coding interview is a **45-minute** data-structures-and-algorithms round. You typically solve **2 challenges in about 40 minutes** from Meta’s question bank. If time runs short, you may explain the approach to the final challenge without coding it. Choose your programming language at the start, make each solution work before optimizing, and leave time for follow-up constraints.

Expect medium to medium-hard questions, including modified familiar problems. These rounds do not ask dynamic-programming questions according to this guide; prioritize trees, graphs, heaps, and string parsing. Read carefully for changes from the standard problem. Prioritize our [most frequently reported Meta questions](https://www.aced.io/questions?company=meta-facebook&type=coding).

Interviewers look for problem solving, coding, verification and debugging, pace, and communication of assumptions, trade-offs, and complexity.

### Recently asked questions

- [Count mismatched brackets](https://www.aced.io/questions/4214/balance-parentheses-minimum-removals) containing round, square, and curly brackets.
- Given a tree and 2 nodes, [find their first shared ancestor](https://www.aced.io/questions/4538/binary-tree-first-common-ancestor), including whether it can be done without traversing to the root.
- [Track the k largest elements](https://www.aced.io/questions/4673/coding-interview-find-kth-largest-element-array) in a running list.
- [Group binary-tree nodes by vertical column](https://www.aced.io/questions/5462/vertical-traversal-of-a-binary-tree).
- Return the index of the largest number, choosing randomly among duplicate maxima.
- Resolve a directory-command string using periods, double periods, and tildes.
- Given a sorted array and element `k`, find the next element greater than `k`.
- [Decide whether a string can become a palindrome after deleting at most one character](https://www.aced.io/questions/6033/valid-palindrome-ii).

## Meta AI-enabled coding interview

The AI-enabled, or AI-assisted, coding round is a **60-minute** MLE interview in a multi-file CoderPad project with an editor-integrated AI assistant. Most generalist loops substitute it for one standard coding round; some use 2 standard rounds or 2 AI-enabled rounds.

The interviewer explains the challenge and CoderPad layout, including a model picker for several LLMs. Expect short files, test files, one core function to implement, and helper functions that may not be needed. The task commonly has 3 phases:

1. Fix a bug.
2. Implement the core logic.
3. Optimize against larger test data.

Prioritize a correct core implementation; completing the final optimization phase is not always required to pass. The round supports **9 languages**, including Python, Java, C++, and TypeScript. Confirm your language with your recruiter and practice in the same environment.

Meta expects you to use the assistant, but does not score prompting skill by itself. Use it for well-defined subtasks such as unfamiliar syntax or a known data structure. Plan the algorithm yourself and treat generated output as a draft. Run test files one at a time, inspect the failing line and function contract, and then prompt the assistant.

Interviewers look for codebase navigation, AI judgment, code reuse, complexity awareness, and clear reasoning about failures and fixes. Check every assistant-referenced function against the codebase before running it.

### Recently asked questions

- Implement the core function in a multi-file codebase so every provided test file passes.
- Given a list of words, return words containing another list word as a substring.
- Extend a solution that works for alphabetic input to numeric and mixed alphanumeric input.
- State the solution’s time complexity and make it more efficient.

Confirm input types in every test file. A bitmask designed for letters can collide when numeric characters appear.

## ML coding interview (AI coding)

On specialized tracks, recruiters may call the ML coding interview “AI coding.” It focuses on implementing data processing and mathematics behind ML systems. Expect **2 or 3 challenges** in one session, each requiring efficient implementation of a core ML operation.

This format is distinct from the generalist AI-enabled coding round despite the shared label. Ask your recruiter which format you will take and whether an AI assistant is available. Practice linear algebra and nearest-neighbor operations from scratch, especially with sparse inputs and time complexity in mind. Relevant building blocks include k-means clustering, nearest-neighbor lookups, and the forward pass of multi-head attention.

Interviewers look for efficient implementation, suitable data structures for sparse or high-dimensional data, mathematical fluency, pace, and communication of complexity and trade-offs.

### Recently asked questions

- Implement sparse matrix-vector multiplication efficiently.
- [Implement a k-nearest-neighbor search](https://www.aced.io/questions/3489/implement-k-nearest-neighbors-algorithm).
- [Design a data structure for a sparse vector](https://www.aced.io/questions/1213/design-data-structure-sparse-vector) that supports efficient operations.

## Behavioral interview

The behavioral interview is a **45-minute** conversation about using AI to drive impact, continued learning, partnering, ambiguity, and communication. Prepare **4–5 stories**, including at least 2 in which you used AI to ship something and caught an AI mistake.

Expect detailed follow-ups on your most significant project: why you chose it, its impact, and how you measured that impact. Tell conflict stories end to end, including the outcome. At E6, a weak behavioral round can result in a down-leveled offer, so match your stories to the scope of the target level.

Interviewers look for AI-driven impact, continuous AI learning, partnering, embracing ambiguity, effective communication, and measured impact tied to business outcomes.

### Recently asked questions

- Tell me about a time you were under schedule pressure to deliver something you had committed to. What did you do?
- Tell me about a time you received a [less-than-stellar review from your manager](https://www.aced.io/questions/995/example-tough-feedback-received). How did you grow from it?
- How do you push development forward without explicit support from the rest of your team?
- Tell me about a time you had to pivot unexpectedly. How did you handle not controlling the direction of your work?
- Tell me about a time you worked independently.
- How have you used AI to increase your impact?
- How do you safely use AI in day-to-day work?

## Meta ML system design interview

The ML system design round is roughly **45 minutes** and open-ended. You design a production ML system end to end, most often a recommendation or ranking system. Some rounds may focus on generative AI or LLM systems, so prepare an LLM design such as a retrieval-augmented generation (RAG) feature alongside ranking designs.

Set your structure at the beginning. Define the problem, align business and ML objectives, cover data and features at a high level, and reserve the most time for model choice and evaluation. Avoid theory and mathematical derivations unless requested. Practice drawing diagrams in Mermaid because Meta conducts design interviews in CoderPad with the same AI assistant available in coding rounds.

Interviewers look for problem framing, modeling choices, data and features, evaluation and deployment, and a clear rationale for every decision.

### Recently asked questions

- Design a recommendation system ranking locations in a feed for a user who just arrived in a new city; explain how you would featurize the ranking data.
- [Design an Instagram ranking model](https://www.aced.io/questions/4012/design-instagram-ranking-model).
- Design recommendation and ranking for Instagram Reels.
- Design a harmful-content detection system.
- Design comment ranking on Facebook posts.

## AI system design and system design interviews: specialized tracks

Specialized loops such as computer vision can include **2 design interviews of about 45 minutes each**: an AI system design round and a full system design round.

The AI system design, also called ML modeling, is a back-and-forth domain problem. Defend modeling trade-offs and adapt when challenged. The full system design round asks for a complete product system balancing the algorithm, data collection, and production handling. Keep algorithm details concise unless requested and budget time across all 3 areas.

Interviewers look for modeling approach, iteration, data collection, production handling, and system trade-offs.

### Recently asked questions

- Estimate depth from a single image.
- [Design a system like Google Street View](https://www.aced.io/questions/2109/design-google-street-view-1000-cars).
- [Estimate Google Maps Street View storage](https://www.aced.io/questions/1868/google-maps-street-view-storage).

## How to prepare

1. **Confirm your track early.** Ask whether the loop includes AI-enabled coding or ML coding and whether it includes one design round or 2.
2. **Practice at Meta’s pace.** Solve 2 challenges back to back in 40 minutes, sometimes adding a third, and review heap patterns such as running top-k.
3. **Prepare every project deeply.** Know motivation, assumptions, trade-offs, limitations, and validation. Project questions occur in pre-screen, behavioral, and design rounds.
4. **Think out loud.** Narrate reasoning, ask clarifying questions, and maintain structure.
5. **Review ML fundamentals at application level.** Connect concepts to implementation and design choices.
6. **Practice both coding environments.** Use a plain editor without autocomplete for standard solutions and an editor with a built-in AI assistant for separate practice sessions.
7. **Start ML design preparation early.** Begin with recommendation and ranking designs, then expand to other systems.
8. **Use mock interviews.** Run timed coding and ML design sessions with [peer or AI mock interviews](https://www.aced.io/practice). For targeted ML-design and behavioral feedback, [book an expert coach](https://www.aced.io/coaching).

## Additional resources

- [Machine Learning Engineer Interview Course](https://www.aced.io/courses/ml-engineer)
- [ML System Design Interview Course](https://www.aced.io/courses/ml-system-design)
- [ML Coding Interview Course](https://www.aced.io/courses/ml-coding)
- [Software Engineering Interview Course](https://www.aced.io/courses/software-engineering)
- [Engineering Behavioral Interview Course](https://www.aced.io/courses/eng-behavioral)
- [Meta MLE interview questions](https://www.aced.io/questions?company=meta-facebook&role=ml-engineer)
- [Meta interview experiences](https://www.aced.io/experiences?company=meta-facebook)
- [Get a Job at Meta: Interview Process and Top Questions](https://www.aced.io/blog/the-meta-facebook-interview-process)

## FAQs

### What is the AI coding interview at Meta?

“AI coding” refers to 2 different interviews depending on track. Generalist loops may include a 60-minute AI-enabled round debugging and extending a multi-file codebase with an integrated assistant. Specialized loops such as computer vision may include ML coding focused on operations such as sparse matrix-vector multiplication and k-nearest-neighbor search.

### How many rounds are in the final loop?

The final loop has **4–5 interviews**, depending on track. Generalist loops typically include standard coding, AI-enabled coding, ML system design, and behavioral rounds. Specialized loops may replace AI-enabled coding with ML coding and add a second design round; E6 candidates often receive 2 ML system design rounds.

### Is the process team-specific?

The process is team-independent through the final loop, with team matching after passing. The applied track—such as generalist MLE or computer vision—still determines coding and design rounds.

### How long does the process take?

Allow roughly **4–8 weeks** from first contact to a hiring decision. Results often arrive within a week of the loop. Team matching may add several more weeks depending on location and headcount. Share competing-offer deadlines with your recruiter and follow up if you have not heard back within 2 weeks of the loop.

### How much does a Meta machine learning engineer make?

The guide reports these compensation figures by level, citing [Levels.fyi](https://www.levels.fyi/companies/meta/salaries/software-engineer/title/machine-learning-engineer):

- **E3 (Entry-level MLE):** ~$187K
- **E4 (MLE):** ~$310K
- **E5 (Senior MLE):** ~$458K
- **E6 (Staff MLE):** ~$782K

Stock makes up most total pay at E6 and is nearly double base salary there. Base salary rises from **$138K at E3** to **$261K at E6**.

## Aced interview preparation

We provide free interview guides, insider tips, and courses through Aced, formerly Exponent. [Create a free account](https://www.aced.io/signup?src=guide).

---

[Read this page’s summary](https://www.aced.io/agents/api/public/v1/page-tldr?url=https%3A%2F%2Fwww.aced.io%2Fguides%2Fmeta-machine-learning-engineer-interview&trace=t_679ff9ab98f3)

## Ask a follow-up about this page

Need a more specific answer? Ask the exact question you have. It returns a fresh, cited answer from Aced (formerly Exponent)’s knowledge base. No account or API key required.

Open this URL with your own question, URL-encoded. Include your question in q. All research inputs are optional; no account or setup.

`GET https://www.aced.io/agents/agent/faq?q={question}&trace=t_679ff9ab98f3`

[Tool guide and examples](https://www.aced.io/agents/agent-access) · [OpenAPI](https://www.aced.io/agents/api/public/v1/openapi.json) · [Page TLDR form](https://www.aced.io/agents/agent-tools/page-tldr?url=https%3A%2F%2Fwww.aced.io%2Fguides%2Fmeta-machine-learning-engineer-interview). Opening a form does not execute a tool.