← Meta Interview Insights

Meta·Data Scientist·Technical Phone Screen·Senior

Senior
Jun 2026

Summary

Tough technical screen for a DS role at Meta. The whole thing was basically one long Bayesian stats problem with a bunch of sub-parts, and they pushed hard on the practical implications, not just the math.

Questions Asked (4)

Q1

You have a binary KPI with Beta(1,1) priors and 10,000 users per arm. One arm gets 515 conversions, the other 500. Compute the posterior for each arm and the probability that lift is greater than zero.

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

The Beta(1,1) prior part is easy since it's just uniform, so the posteriors are Beta(516,9486) and Beta(501,9501).

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

First, recognize that with Beta(1,1) priors and binomial data, the posteriors are Beta(1+conversions, 1+non-conversions) for each arm. Then, compute the posterior distributions and estimate the probability that the difference in conversion rates (lift) is greater than zero, either analytically or via Monte Carlo simulation.

Pro tip: In practice, use Monte Carlo simulation to estimate the probability that lift > 0, as it's straightforward and scales to more complex metrics. Also, mention that with large sample sizes, the posterior is approximately normal, so you can use a normal approximation for a quick sanity check.

1. Identify priors and likelihood

Note that Beta(1,1) is a uniform prior, and the data follows a binomial distribution. The posterior for each arm is Beta(α + conversions, β + non-conversions).

2. Compute posterior parameters

For arm A (515 conversions): posterior is Beta(1+515, 1+9485) = Beta(516, 9486). For arm B (500 conversions): posterior is Beta(1+500, 1+9500) = Beta(501, 9501).

3. Estimate probability lift > 0

Use Monte Carlo simulation: draw many samples from each posterior, compute the difference (or ratio) of conversion rates, and calculate the proportion of samples where the difference is positive.

4. Interpret and communicate results

Report the probability that lift > 0, and optionally provide a credible interval for the lift. Discuss practical significance and any assumptions.

Key Points to Mention

  • Beta-Binomial conjugacy: posterior is Beta(α + successes, β + failures).
  • Posterior parameters: Arm A ~ Beta(516, 9486), Arm B ~ Beta(501, 9501).
  • Monte Carlo simulation for estimating P(lift > 0).
  • Normal approximation for large samples: difference in proportions is approximately normal.
  • Consideration of practical significance vs. statistical significance.
  • Assumptions: independent users, no interference, prior choice.

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

Q2

How would you define a decision rule using expected loss when the costs of false positives and false negatives are asymmetric?

A/B Testing & ExperimentationTechnical Trade-offs
Author's notes

This is where I actually felt okay.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by defining expected loss as the weighted sum of false positive and false negative costs, then derive the optimal decision threshold that minimizes this loss. Explain how the threshold shifts based on the cost ratio, and discuss practical implications for model deployment and evaluation.

Pro tip: Emphasize that in real-world applications, the cost ratio is often estimated from business context, and you should validate the chosen threshold using a holdout set or cross-validation to ensure robustness.

1. Define the cost structure

Clearly state the costs associated with false positives (C_FP) and false negatives (C_FN), and note that these are asymmetric. Explain that these costs can be monetary, time, or other business metrics.

2. Formulate expected loss

Express the expected loss for a given decision threshold as: E[Loss] = C_FP * P(FP) + C_FN * P(FN), where P(FP) and P(FN) depend on the threshold and the model's score distribution.

3. Derive the optimal threshold

Show that the optimal threshold is where the likelihood ratio equals the cost ratio: P(x|positive)/P(x|negative) = C_FP / C_FN. Alternatively, set the threshold such that the posterior probability of positive class equals C_FP / (C_FP + C_FN).

4. Interpret and apply

Explain how the threshold shifts: if false negatives are more costly, lower the threshold to predict positive more often. Discuss how to estimate costs and validate the threshold on validation data.

Key Points to Mention

  • Expected loss formula: E[Loss] = C_FP * P(FP) + C_FN * P(FN)
  • Optimal threshold condition: P(y=1|x) = C_FP / (C_FP + C_FN)
  • Cost ratio determines the trade-off between false positives and false negatives
  • Threshold can be tuned post-training without retraining the model
  • Use validation set to estimate P(FP) and P(FN) at different thresholds
  • Consider business impact and communicate cost assumptions to stakeholders

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

Q3

Compare this Bayesian decision approach to a frequentist test run under optional stopping. What are the operational differences, and how would you communicate posterior probabilities versus p-values to stakeholders?

A/B Testing & ExperimentationStakeholder Management
Author's notes

Optional stopping is where frequentist tests fall apart and I said so pretty directly.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by contrasting the philosophical and practical implications of Bayesian and frequentist approaches under optional stopping, emphasizing that Bayesian inference remains valid while frequentist p-values become miscalibrated. Then, discuss how to translate posterior probabilities into actionable insights for stakeholders, avoiding technical jargon and focusing on decision-making.

Pro tip: Frame the Bayesian approach as a tool for continuous monitoring and decision-making, which aligns with business needs, while cautioning that frequentist methods require pre-specified stopping rules to avoid inflated error rates. This shows you understand both the statistical rigor and the practical trade-offs.

1. Define the core difference

Explain that Bayesian inference updates beliefs continuously and is not affected by optional stopping, whereas frequentist p-values assume a fixed sample size and become invalid if you peek at data repeatedly.

2. Operational implications

Describe how optional stopping leads to inflated Type I error in frequentist tests, while Bayesian posterior probabilities remain coherent. Mention that Bayesian methods allow for more flexible experimentation but require prior specification.

3. Communicating to stakeholders

For posterior probabilities, use plain language like 'There is an 85% chance that variant B is better' and tie it to expected loss or gain. For p-values, explain them as 'the probability of seeing this result if there were no effect' and highlight their limitations under peeking.

4. Decision-making focus

Emphasize that Bayesian posterior probabilities directly answer the question stakeholders care about: 'What is the probability that B is better?' while p-values answer a different, less intuitive question.

5. Recommendation and trade-offs

Suggest that for Meta's scale, Bayesian methods can enable faster, more intuitive decisions, but acknowledge that frequentist methods are simpler and more familiar. Recommend a hybrid or clear guidelines depending on the context.

Key Points to Mention

  • Optional stopping: Bayesian inference is not affected by multiple looks, while frequentist p-values are.
  • Type I error inflation in frequentist tests under optional stopping.
  • Posterior probabilities are conditional on data and prior, and can be directly interpreted as belief.
  • P-values are not the probability that the null hypothesis is true; they are the probability of data given null.
  • Stakeholder communication: use decision-focused language, avoid technical jargon, and relate to business metrics.
  • Practical considerations: prior selection, computational cost, and organizational familiarity.

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

Q4

How would you set or elicit priors in a way that avoids over-optimism in your experiments?

A/B Testing & ExperimentationCross-functional Alignment
Author's notes

Blanked for a second.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by acknowledging that over-optimism often stems from biased prior elicitation, such as relying on stakeholder opinions or historical data without adjustment. Then, describe a structured process that combines multiple sources of evidence, uses conservative adjustments, and validates priors through calibration. Emphasize the importance of pre-registration and sensitivity analysis to detect and mitigate over-optimism.

Pro tip: Frame priors as a range rather than a point estimate, and explicitly discuss how you would detect and correct for over-optimism if early experiment results deviate from prior expectations.

1. Identify and diversify prior sources

Gather priors from multiple independent sources: historical experiments, domain expertise, and analogous studies. Avoid relying solely on optimistic stakeholder projections.

2. Apply conservative adjustments

Systematically shrink overly optimistic priors toward more neutral values using techniques like empirical Bayes or discounting factors based on past miscalibration.

3. Pre-register and stress-test priors

Document priors before the experiment and conduct sensitivity analyses to see how results change under different prior assumptions, including pessimistic scenarios.

4. Validate and recalibrate

After the experiment, compare prior predictions with observed outcomes to assess calibration. Use this feedback to refine future prior elicitation processes.

Key Points to Mention

  • Bayesian priors and the risk of over-optimism from biased elicitation
  • Empirical Bayes and hierarchical models to borrow strength while adjusting for optimism
  • Pre-registration of analysis plans to prevent post-hoc rationalization
  • Sensitivity analysis and prior predictive checks to test robustness
  • Calibration metrics (e.g., coverage of credible intervals) to evaluate prior quality
  • Cross-functional collaboration to align on realistic assumptions and avoid groupthink

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