##### Scenario You are a data scientist at Meta. Fake accounts (bots, spam, scams, impersonation, coordinated inauthentic behavior, and compromised legitimate accounts) are a persistent problem across the platform's social and social-commerce surfaces. ##### Question Design an end-to-end machine learning system to identify fake accounts in the user base, then reason about evaluation, operations, and business impact: 1. **Approach and modeling.** Describe a full approach to identify fake accounts. Outline how you would build and train a model, including the features you would engineer and whether you would use supervised, unsupervised, or semi/weakly-supervised methods. 2. **Evaluation metrics.** Which evaluation metrics would you monitor, and how would you prioritize them given the precision–recall trade-off and the business cost of false positives vs. false negatives? 3. **Drift monitoring.** Once deployed, how would you monitor the model for data drift and concept/performance drift, and how would you decide when to retrain? 4. **Ecosystem impact.** What is the impact of fake users on the overall ecosystem — buyers, sellers, and ads/measurement? 5. **Interpreting a finding.** Suppose an analysis shows that 3% of accounts are fake. What conclusions or next steps does this number suggest? ##### Hints Discuss labeling and class imbalance, precision–recall trade-offs, supervised vs. unsupervised methods, cost-sensitive thresholds tied to enforcement actions, graph/network signals, A/B testing for business impact, and how account-level prevalence can differ from activity-level prevalence.

Identify Fake Accounts Using Machine Learning Techniques evaluates core ML concepts, assumptions, math intuition, training/evaluation trade-offs, and practical failure modes in a realistic interview setting. A strong answer states assumptions, handles edge cases, explains trade-offs, and shows how to validate the result clearly.

How do I approach Machine Learning interview questions?

Machine Learning questions require understanding of core concepts and practice. PracHub provides solutions with explanations to help you master machine learning interviews.

What difficulty level is this interview question?

This is a medium difficulty Machine Learning question, commonly asked during Onsite rounds at Meta.

What role is this question designed for?

This question is commonly asked for Data Scientist candidates at Meta during technical interviews.

Identify Fake Accounts Using Machine Learning Techniques

Scenario

You are a data scientist at Meta. Fake accounts (bots, spam, scams, impersonation, coordinated inauthentic behavior, and compromised legitimate accounts) are a persistent problem across the platform's social and social-commerce surfaces.

Question

Design an end-to-end machine learning system to identify fake accounts in the user base, then reason about evaluation, operations, and business impact:

Approach and modeling. Describe a full approach to identify fake accounts. Outline how you would build and train a model, including the features you would engineer and whether you would use supervised, unsupervised, or semi/weakly-supervised methods.
Evaluation metrics. Which evaluation metrics would you monitor, and how would you prioritize them given the precision–recall trade-off and the business cost of false positives vs. false negatives?
Drift monitoring. Once deployed, how would you monitor the model for data drift and concept/performance drift, and how would you decide when to retrain?
Ecosystem impact. What is the impact of fake users on the overall ecosystem — buyers, sellers, and ads/measurement?
Interpreting a finding. Suppose an analysis shows that 3% of accounts are fake. What conclusions or next steps does this number suggest?

Hints

Discuss labeling and class imbalance, precision–recall trade-offs, supervised vs. unsupervised methods, cost-sensitive thresholds tied to enforcement actions, graph/network signals, A/B testing for business impact, and how account-level prevalence can differ from activity-level prevalence.

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

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

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

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?