Detect and Reduce Spammy Friend Requests Effectively

Quick Overview

This interview question evaluates core ML concepts, assumptions, math intuition, training/evaluation trade-offs, and practical failure modes in a realistic interview setting. A strong answer for Detect and Reduce Spammy Friend Requests Effectively states assumptions, handles edge cases, explains trade-offs, and shows how to validate the result clearly.

Detect and Reduce Spammy Friend Requests Effectively

Company: Meta

Role: Data Scientist

Category: Machine Learning

Difficulty: medium

Interview Round: Technical Screen

##### Scenario A social media platform wants to detect and reduce spammy friend-requests in order to protect user experience. ##### Question How would you define a spammy friend request on our platform? What key features or signals would you engineer to identify such requests? Describe how you would build a classification model for this task. If no labeled data exist, what strategies would you use to obtain or generate labels? Once the model is live, how would you use it to improve the overall user experience? How do you determine and monitor the appropriate precision-recall trade-off for this problem? ##### Hints Discuss heuristic labeling, human-in-the-loop, semi-supervised learning, A/B tests, threshold tuning, and business impact of false positives vs. false negatives.

Quick Answer: This interview question evaluates core ML concepts, assumptions, math intuition, training/evaluation trade-offs, and practical failure modes in a realistic interview setting. A strong answer for Detect and Reduce Spammy Friend Requests Effectively states assumptions, handles edge cases, explains trade-offs, and shows how to validate the result clearly.

|Home/Machine Learning/Meta
Meta logo
Meta
Aug 4, 2025, 10:55 AM
mediumData ScientistTechnical ScreenMachine Learning
2
0

Detect and Reduce Spammy Friend Requests Effectively

Detecting Spammy Friend Requests

Context

Assume a consumer social platform where users can send friend requests (optionally with a short message). The goal is to protect user experience by reducing spammy requests while preserving legitimate connections, especially for new users.

Tasks

  1. Definition
    • Clearly define what constitutes a "spammy" friend request on this platform.
  2. Feature Engineering
    • List the key features/signals you would engineer to identify spammy requests (sender, recipient, pairwise, graph, content, device, temporal, and feedback signals).
  3. Modeling Approach
    • Describe how you would build an initial classification model, including data splitting, handling class imbalance, model choice, interpretability, and calibration.
  4. Labels When None Exist
    • If no labeled data exist, explain strategies to obtain/generate labels (heuristics, human-in-the-loop, semi-/weak supervision, and positive–unlabeled learning).
  5. Product Integration
    • Once live, how would you use the model to improve user experience (interventions and experimentation)?
  6. Precision–Recall Trade-off
    • How would you determine, set, and monitor the appropriate precision–recall trade-off for this problem?

Constraints & Assumptions

  • Preserve the scope, facts, inputs, and requested outputs from the prompt above.
  • If the prompt leaves a detail unspecified, state a reasonable assumption before relying on it.
  • Keep the answer interview-ready: concise enough to present, but concrete enough to implement or evaluate.

Clarifying Questions to Ask Guidance

  • Clarify the task, data shape, labels, constraints, and evaluation metric.
  • State assumptions behind the math or modeling technique you choose.
  • Connect theory to practical training, debugging, and deployment implications.

What a Strong Answer Covers Guidance

  • Correct definitions and formulas where the prompt requires them.
  • A practical explanation of how the method behaves on real data.
  • Trade-offs, failure modes, diagnostics, and mitigation strategies.
  • Evaluation choices that match the product or modeling objective.

Follow-up Questions Guidance

  • How would noisy labels, class imbalance, or distribution shift affect the answer?
  • What would you monitor after deployment?
  • Which baseline would you compare against first?
Loading comments...

Browse More Questions

More Machine Learning•More Meta•More Data Scientist•Meta Data Scientist•Meta Machine Learning•Data Scientist Machine Learning