Meta Generative AI Engineer Interview Questions (2026)

1. 1Recruiter screen30 minEasy

   Process overview, level calibration, and prep guidance — Meta recruiters actively coach on round formats.

2. 2Technical screen45 minHard

   Two DSA problems in 45 minutes on a plain shared editor with no autocomplete or execution.

3. 3Coding round ('Ninja')45 minHard

   Two more problems at loop difficulty; clean near-compilable code and verbalized complexity analysis expected.

4. 4System design ('Pirate')45 minHard

   Design a Meta-scale product system (feed, Stories, chat) with emphasis on read-heavy fan-out, caching, and data modeling.

5. 5Behavioral ('Jedi')45 minMedium

   Deep past-experience discussion on conflict, growth, and impact aligned to Meta values; graded as a real signal round.

1. Q1

   At Meta, you are interviewing for the Integrity/Evals/Agents Engineer loop and working on a messaging assistant with high daily traffic. When should you use a deterministic workflow instead of an autonomous agent?

   FoundationalAgentic product designAgents

   How to answer:

   Use workflows when steps are known, compliance matters, and determinism is valuable. Use agentic behavior for ambiguous tasks requiring planning, but bound it with state machines, budgets, and approvals. In this Meta loop, I would explicitly connect the decision to pragmatic execution, product scale, ranking/relevance, measurable impact, efficiency, and integrity/safety trade-offs.

2. Q2

   At Meta, you are interviewing for the Integrity/Evals/Agents Engineer loop and working on a retrieval feature for help-center answers. Design memory for a personal assistant agent without creating privacy risks

   FoundationalCoding screenAgents

   How to answer:

   Separate short-term session state from long-term user preferences, require user-visible controls, store minimal facts with provenance, expire stale data, and avoid storing sensitive information by default. In this Meta loop, I would explicitly connect the decision to pragmatic execution, product scale, ranking/relevance, measurable impact, efficiency, and integrity/safety trade-offs.

3. Q3

   At Meta, you are interviewing for the Integrity/Evals/Agents Engineer loop and working on a retrieval feature for help-center answers. What state should be persisted between agent steps?

   IntermediateCoding screenAgents

   How to answer:

   Persist user goal, plan, tool calls/results, decisions, intermediate artifacts, permissions, budgets, and trace IDs. Avoid persisting raw sensitive data unless necessary and approved. In this Meta loop, I would explicitly connect the decision to pragmatic execution, product scale, ranking/relevance, measurable impact, efficiency, and integrity/safety trade-offs.

4. Q4

   At Meta, you are interviewing for the Integrity/Evals/Agents Engineer loop and working on a GenAI tool that must be efficient on mobile. An agent retries a failing tool and creates duplicate tickets. What went wrong?

   IntermediateBehavioral / executionAgents

   How to answer:

   The tool call was not idempotent or lacked a deduplication key. Add idempotency tokens, retry policies by error class, confirmation for side effects, and reconciliation checks. In this Meta loop, I would explicitly connect the decision to pragmatic execution, product scale, ranking/relevance, measurable impact, efficiency, and integrity/safety trade-offs.

5. Q5

   At Meta, you are interviewing for the Integrity/Evals/Agents Engineer loop and working on a messaging assistant with high daily traffic. How do you evaluate whether an agent completed a task correctly?

   StaffAgentic product designAgents

   How to answer:

   Use scenario-based evals with initial state, allowed actions, hidden expected outcomes, tool traces, final artifacts, and human/automated grading. Score task success, safety, cost, and efficiency. In this Meta loop, I would explicitly connect the decision to pragmatic execution, product scale, ranking/relevance, measurable impact, efficiency, and integrity/safety trade-offs.

6. Q6

   At Meta, you are interviewing for the Integrity/Evals/Agents Engineer loop and working on a content integrity workflow. An agent can read documents and send emails. What permission model do you implement?

   StaffIntegrity, evals, and measurementAgents

   How to answer:

   Use least privilege, separate read and write scopes, per-tool consent, tenant/user ACL checks, action previews, audit logs, and revocation. The model should never be the authority on permissions. In this Meta loop, I would explicitly connect the decision to pragmatic execution, product scale, ranking/relevance, measurable impact, efficiency, and integrity/safety trade-offs.

7. Q7

   At Meta, you are interviewing for the AI Infra/Retrieval Engineer loop and working on a messaging assistant with high daily traffic. How do you choose an embedding model for a semantic search system?

   FoundationalRetrieval infrastructure designEmbeddings

   How to answer:

   Evaluate on representative queries/documents, languages, domain terms, latency, cost, dimensionality, and compatibility with the vector index. Avoid relying only on public benchmarks. In this Meta loop, I would explicitly connect the decision to pragmatic execution, product scale, ranking/relevance, measurable impact, efficiency, and integrity/safety trade-offs.

8. Q8

   At Meta, you are interviewing for the AI Infra/Retrieval Engineer loop and working on a content integrity workflow. Your embeddings work for general docs but fail on legal contracts. Why?

   FoundationalReliability and scalingEmbeddings

   How to answer:

   Legal language has specialized terminology, structure, and semantic distinctions. Improve chunking, domain-specific examples/evals, hybrid search, reranking, or fine-tune/adapt embeddings if available. In this Meta loop, I would explicitly connect the decision to pragmatic execution, product scale, ranking/relevance, measurable impact, efficiency, and integrity/safety trade-offs.

9. Q9

   At Meta, you are interviewing for the AI Infra/Retrieval Engineer loop and working on a GenAI tool that must be efficient on mobile. How would you embed 100 million chunks cost-effectively?

   IntermediateBehavioral / executionEmbeddings

   How to answer:

   Use batch jobs, rate-limit handling, retries, deduplication, incremental checkpoints, cost estimates, monitoring, and idempotent writes. Avoid re-embedding unchanged chunks. In this Meta loop, I would explicitly connect the decision to pragmatic execution, product scale, ranking/relevance, measurable impact, efficiency, and integrity/safety trade-offs.

10. Q10

    At Meta, you are interviewing for the AI Infra/Retrieval Engineer loop and working on an AI assistant for creators. What are hard negatives and why do they matter for retrieval evaluation?

    IntermediateRetrieval infrastructure designEmbeddings

    How to answer:

    Hard negatives are irrelevant items that look semantically similar to the query. They test whether retrieval distinguishes subtle differences and improve training/evaluation rigor. In this Meta loop, I would explicitly connect the decision to pragmatic execution, product scale, ranking/relevance, measurable impact, efficiency, and integrity/safety trade-offs.

11. Q11

    At Meta, you are interviewing for the AI Infra/Retrieval Engineer loop and working on a retrieval feature for help-center answers. How would you use embeddings to cluster user feedback?

    IntermediateCoding screenEmbeddings

    How to answer:

    Generate embeddings, reduce or cluster them, label clusters with representative examples, and track cluster size over time. Validate clusters manually before making product decisions. In this Meta loop, I would explicitly connect the decision to pragmatic execution, product scale, ranking/relevance, measurable impact, efficiency, and integrity/safety trade-offs.

12. Q12

    At Meta, you are interviewing for the AI Infra/Retrieval Engineer loop and working on a GenAI tool that must be efficient on mobile. Should you embed queries and documents the same way?

    IntermediateBehavioral / executionEmbeddings

    How to answer:

    Often yes for symmetric models, but some systems use distinct query/document encoders optimized for retrieval direction. Follow the model's intended usage and evaluate both. In this Meta loop, I would explicitly connect the decision to pragmatic execution, product scale, ranking/relevance, measurable impact, efficiency, and integrity/safety trade-offs.

13. Q13

    At Meta, you are interviewing for the AI Infra/Retrieval Engineer loop and working on a messaging assistant with high daily traffic. Why might embeddings miss exact product codes or error IDs?

    SeniorRetrieval infrastructure designEmbeddings

    How to answer:

    Dense embeddings capture semantic similarity and may not preserve rare exact strings. Hybrid retrieval with keyword/BM25 or metadata filters improves exact-match cases. In this Meta loop, I would explicitly connect the decision to pragmatic execution, product scale, ranking/relevance, measurable impact, efficiency, and integrity/safety trade-offs.

14. Q14

    At Meta, you are interviewing for the AI Infra/Retrieval Engineer loop and working on a content integrity workflow. What is embedding drift and how do you monitor it?

    SeniorReliability and scalingEmbeddings

    How to answer:

    Drift occurs when corpus/query distribution or embedding model versions change and retrieval quality shifts. Monitor recall evals, nearest-neighbor distributions, query clusters, and model/index version changes. In this Meta loop, I would explicitly connect the decision to pragmatic execution, product scale, ranking/relevance, measurable impact, efficiency, and integrity/safety trade-offs.

15. Q15

    At Meta, you are interviewing for the AI Infra/Retrieval Engineer loop and working on a messaging assistant with high daily traffic. How would you collect labels for embedding retrieval quality?

    StaffRetrieval infrastructure designEmbeddings

    How to answer:

    Sample real queries, label relevant documents/chunks, include hard negatives and no-answer cases, and measure inter-annotator agreement. Convert production failures into eval examples. In this Meta loop, I would explicitly connect the decision to pragmatic execution, product scale, ranking/relevance, measurable impact, efficiency, and integrity/safety trade-offs.

Practice these with instant AI feedback in a live mock interview → Start a Meta Generative AI Engineer mock

Be fast and correct on coding; for design, drive the conversation; prepare impact-focused behavioral stories

Indicative Generative AI Engineer pay in India: ~₹14–60 LPA (role-level range, not a Meta-specific figure).

Other roles at Meta

Generative AI Engineer interviews at other companies