← Meta Interview Insights

Meta·Data Scientist·Technical Phone Screen·Senior

SeniorPrefer not to say
Apr 2026

Summary

Meta data scientist interview with a deep dive into a single metric (comments per daily active user) across six sub-parts covering distributions, bootstrapping, scaling, summary stats, data types, and estimation. The whole thing felt like a stats exam more than a conversation, and some parts I handled better than others.

Questions Asked (6)

Q1

What distribution would you expect for comments per daily active user? Is it discrete or continuous, and why might Poisson be a poor fit?

Product Analytics & MetricsTechnical Trade-offs
Author's notes

This part I actually felt okay about.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying that comments per daily active user is a discrete count variable, then describe the expected distribution as right-skewed with a long tail, likely overdispersed. Explain why Poisson is a poor fit due to its equidispersion assumption and inability to capture excess zeros and heavy tails, and suggest alternatives like negative binomial or zero-inflated models.

Pro tip: Mention that in practice, you'd validate the distribution by plotting the empirical histogram and computing the variance-to-mean ratio; if it's substantially greater than 1, Poisson is inappropriate. Also, note that user heterogeneity and behavioral clustering often lead to overdispersion.

1. Identify variable type

State that comments per daily active user is a count variable, hence discrete, not continuous. It can only take non-negative integer values.

2. Describe expected distribution shape

Explain that the distribution is typically right-skewed: most users comment zero or few times, while a small fraction of highly active users comment many times, creating a long right tail.

3. Discuss Poisson assumptions and limitations

Poisson assumes mean equals variance (equidispersion) and independence of events. In reality, comments per user often show overdispersion (variance > mean) due to user heterogeneity and clustering, making Poisson a poor fit.

4. Propose alternative distributions

Suggest negative binomial (accounts for overdispersion), zero-inflated Poisson or negative binomial (if excess zeros), or a mixture model to capture user-level heterogeneity.

5. Mention validation and implications

Emphasize checking the variance-to-mean ratio and plotting the empirical distribution. The choice of distribution affects modeling, inference, and product decisions (e.g., identifying power users).

Key Points to Mention

  • Discrete vs continuous: comments per user is a count, so discrete.
  • Right-skewed distribution with a long tail due to heavy users.
  • Poisson assumes mean = variance, but real data often overdispersed.
  • Overdispersion arises from user heterogeneity and behavioral clustering.
  • Negative binomial or zero-inflated models as better alternatives.
  • Practical validation: variance-to-mean ratio and empirical histogram.

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.

Q2

You resample 10,000 users with replacement 100,000 times and compute the sample mean each time. What does that bootstrap distribution look like, and what determines whether it's approximately normal?

A/B Testing & ExperimentationProduct Analytics & Metrics
Author's notes

I knew the central limit theorem angle but got a bit tangled explaining when the skew persists.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Explain that the bootstrap distribution of the sample mean is the sampling distribution of the mean under resampling with replacement, and its shape is governed by the Central Limit Theorem. Clarify that with 10,000 users, the distribution will be approximately normal regardless of the population distribution, but the approximation improves with larger sample size and is exact if the population is normal.

Pro tip: Emphasize that the bootstrap distribution estimates the sampling distribution of the mean, and its normality depends on the original sample size (10,000), not the number of bootstrap resamples (100,000). The latter only reduces Monte Carlo error.

1. Define the bootstrap distribution

Describe that each bootstrap sample is drawn with replacement from the original 10,000 users, and the mean is computed. The distribution of these 100,000 means is the bootstrap distribution.

2. Identify the center and spread

State that the bootstrap distribution is centered at the original sample mean and its standard error is approximately the sample standard deviation divided by the square root of 10,000.

3. Discuss shape and normality

Explain that by the Central Limit Theorem, the distribution of the sample mean is approximately normal for large sample sizes. With n=10,000, it will be very close to normal even if the population is skewed.

4. Clarify determinants of normality

Highlight that normality depends on the sample size (10,000) and the population distribution. Larger sample sizes and less skewed populations yield better normal approximations.

5. Address the number of resamples

Note that 100,000 resamples only reduce Monte Carlo error, making the bootstrap distribution smoother, but do not affect its underlying shape or normality.

Key Points to Mention

  • Bootstrap distribution approximates the sampling distribution of the mean.
  • Central Limit Theorem ensures approximate normality for large sample sizes.
  • Sample size (10,000) is the key factor for normality, not the number of resamples.
  • The distribution is centered at the original sample mean.
  • Standard error is estimated by the standard deviation of the bootstrap means.
  • Increasing resamples reduces Monte Carlo variability but does not change the shape.

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.

Q3

If you double the sample size from 10,000 to 20,000, how does the width of the bootstrap distribution of the mean change, and why?

A/B Testing & Experimentation
Author's notes

Easiest part of the whole question.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying that the bootstrap distribution of the mean approximates the sampling distribution, so its width reflects the standard error. Then explain that doubling the sample size reduces the standard error by a factor of √2, so the width of the bootstrap distribution decreases by about 29%. Emphasize that this is a direct consequence of the Central Limit Theorem and the √n relationship.

Pro tip: Mention that the bootstrap distribution's spread estimates the sampling variability, not the population variability, so it shrinks with larger n. Also note that while the width decreases, the shape becomes more normal, which is important for inference.

1. Define the bootstrap distribution and its width

Explain that the bootstrap distribution of the mean is generated by resampling with replacement from the original sample, and its width (e.g., standard deviation or interquartile range) estimates the standard error of the mean.

2. Relate width to standard error

State that the standard error of the mean is σ/√n, where σ is the population standard deviation. Thus, the width of the bootstrap distribution is proportional to 1/√n.

3. Calculate the effect of doubling n

When n doubles from 10,000 to 20,000, the standard error changes by a factor of 1/√2 ≈ 0.707. So the width decreases by about 29.3% (or is multiplied by 0.707).

4. Explain the intuition and implications

Discuss that larger samples provide more precise estimates, reducing variability. Mention that the bootstrap distribution becomes narrower and more normal, which improves the accuracy of confidence intervals and hypothesis tests.

Key Points to Mention

  • Bootstrap distribution approximates the sampling distribution of the mean.
  • Standard error of the mean is σ/√n, so width scales with 1/√n.
  • Doubling n reduces the width by a factor of √2 (about 29% decrease).
  • The bootstrap distribution's spread reflects sampling variability, not population variability.
  • Larger n also makes the bootstrap distribution more normal (Central Limit Theorem).
  • This reduction in width leads to narrower confidence intervals and more power.

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.

Q4

Compare mean, median, mode, and p95 for this metric. Which is most stable, which matters most for decisions, and how do you handle p95 for a discrete count variable?

Product Analytics & MetricsData Modeling
Author's notes

The mode being zero is obvious once you think about it but I fumbled the p95 discrete-tie question.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by defining each statistic and its properties, then compare their stability and decision-making relevance in the context of a product metric. Finally, address the specific challenge of computing p95 for a discrete count variable, discussing methods like interpolation or using the nearest-rank method.

Pro tip: Emphasize that the choice of statistic depends on the metric's distribution and the business question—mean for totals, median for typical user experience, p95 for tail risks—and that for discrete counts, p95 should be interpreted as the value below which 95% of observations fall, often using the nearest-rank method to avoid fractional counts.

1. Define and Compare Statistics

Briefly define mean, median, mode, and p95, and explain how each summarizes the data differently. Highlight that mean is sensitive to outliers, median is robust, mode is for categorical or discrete peaks, and p95 captures the tail.

2. Assess Stability

Discuss which statistic is most stable across samples. Typically, the median is more stable than the mean for skewed data, while p95 can be volatile due to tail sensitivity. Mode can be unstable if the distribution is flat or multimodal.

3. Determine Decision Relevance

Explain which statistic matters most for decisions. For product metrics, the mean might drive revenue, but the median reflects typical user experience, and p95 highlights worst-case scenarios for performance or satisfaction. The choice depends on the business goal.

4. Handle p95 for Discrete Counts

Address that p95 for discrete counts can be non-integer. Use methods like nearest-rank (the smallest value with cumulative frequency ≥95%) or interpolation, but be cautious as interpolation may produce impossible values. Discuss the trade-offs and recommend nearest-rank for interpretability.

Key Points to Mention

  • Mean is sensitive to outliers and skew, while median is robust and represents the typical value.
  • Mode is useful for identifying the most common value, especially for discrete or categorical data, but can be misleading for continuous data.
  • p95 is a tail metric that captures the experience of the top 5% of users, important for risk and performance monitoring.
  • Stability: median is generally more stable than mean for skewed distributions; p95 can be unstable with small samples or heavy tails.
  • For decision-making, the mean is often used for overall impact, median for user experience, and p95 for service level objectives (SLOs).
  • For discrete counts, p95 should be computed using the nearest-rank method to avoid fractional counts, and it represents the value below which 95% of observations fall.

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.

Q5

The per-user count is an integer but the mean is a real number. What can go wrong when storing or aggregating these at scale, and how do you keep things numerically stable?

Data ModelingTechnical Trade-offsSystem Design
Author's notes

Truncation versus rounding bias I covered fine.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by distinguishing between integer counts and floating-point means, then discuss the numerical issues that arise during storage and aggregation at scale. Structure your answer around precision loss, overflow, and distributed aggregation challenges, and propose solutions like using higher-precision types, compensated summation, and careful data modeling.

Pro tip: Mention that while integer counts are exact, converting to floating-point for means can introduce rounding errors; using Kahan summation or pairwise summation in distributed settings shows depth. Also, highlight that storing raw counts and computing means on the fly avoids precision loss from pre-aggregated floats.

1. Identify the core issue

Explain that per-user counts are integers (exact) but means are real numbers (approximate), leading to potential precision loss when converting or aggregating.

2. Storage challenges

Discuss how storing means as floats can accumulate rounding errors, especially with many users or high counts, and how integer overflow can occur if counts are stored in fixed-width types.

3. Aggregation challenges

Describe issues in distributed aggregation: summing floats across nodes can lead to non-associativity and precision loss, and merging pre-computed means requires weighted averages to avoid bias.

4. Numerical stability techniques

Propose solutions: use 64-bit integers for counts, compute means from raw counts, employ compensated summation (Kahan) or pairwise summation, and consider using higher-precision floats (e.g., double) or fixed-point arithmetic.

5. Trade-offs and best practices

Discuss trade-offs between precision and performance, and recommend storing raw counts and computing means on demand, or using streaming algorithms for approximate means if exactness is not critical.

Key Points to Mention

  • Integer overflow when counts exceed 2^31 or 2^63
  • Floating-point precision loss (e.g., 0.1 + 0.2 != 0.3)
  • Non-associativity of floating-point addition in distributed aggregation
  • Weighted average for merging means from different partitions
  • Kahan summation or pairwise summation for numerical stability
  • Storing raw counts vs. pre-aggregated means to avoid precision loss

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.

Q6

Given that per-user variance exceeds the mean (overdispersion), write out the standard error of the sample mean and explain when you'd use trimmed means, Winsorization, or a covariate-based variance reduction approach.

A/B Testing & ExperimentationProduct Analytics & MetricsTechnical Trade-offs
Author's notes

The SE formula is straightforward but the discussion around robust estimators versus variance reduction is where it got interesting.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by defining the standard error of the sample mean under overdispersion, clarifying that it is larger than the naive i.i.d. formula due to variance inflation. Then, for each robust method (trimmed means, Winsorization, covariate-based variance reduction), explain when it is appropriate, focusing on trade-offs between bias, efficiency, and interpretability in the context of A/B testing at Meta.

Pro tip: Emphasize that in large-scale A/B tests, the choice of method should be driven by the metric's distribution and business impact; for example, trimmed means are often preferred for heavy-tailed revenue metrics to improve power without introducing excessive bias, but always validate with simulation.

1. Define the standard error under overdispersion

Write the formula for the standard error of the sample mean when per-user variance exceeds the mean, such as SE = sqrt( (σ^2 + φμ) / n ) or using the delta method for ratio metrics. Explain that overdispersion inflates the variance and thus the standard error.

2. Explain trimmed means

Describe trimmed means as removing a fixed percentage of extreme values from both tails. Use when the metric has heavy tails and you want a robust measure of central tendency that reduces the influence of outliers, but be aware of potential bias if the distribution is asymmetric.

3. Explain Winsorization

Describe Winsorization as replacing extreme values with a specified percentile. Use when you want to retain all data points but limit the impact of outliers, often leading to a more efficient estimator than trimming when the tails are not too extreme.

4. Explain covariate-based variance reduction

Describe using pre-experiment covariates (e.g., CUPED) to reduce variance by adjusting for baseline differences. Use when you have relevant pre-period data and want to increase power without altering the metric definition, especially for metrics with high variance.

5. Compare and choose based on context

Discuss trade-offs: trimmed means and Winsorization change the estimand and may introduce bias, while covariate adjustment preserves the estimand but requires valid covariates. Choose based on metric distribution, business goals, and whether the treatment effect is homogeneous.

Key Points to Mention

  • Overdispersion occurs when variance > mean, common in count or revenue metrics, leading to underestimated standard errors if ignored.
  • Standard error formula: SE = sqrt( (σ^2 + φμ) / n ) or more generally sqrt(Var(X)/n) where Var(X) includes overdispersion.
  • Trimmed means: remove a percentage of extremes; robust to outliers but may bias if distribution is skewed.
  • Winsorization: cap extremes at percentiles; retains sample size but can still be influenced by moderate outliers.
  • Covariate-based variance reduction (e.g., CUPED): uses pre-experiment data to reduce variance without changing the metric, increasing power.
  • Trade-offs: bias vs. efficiency, interpretability, and applicability to different metric types (e.g., revenue vs. engagement).

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.