← Uber Interview Insights

Uber·Data Scientist·Technical Phone Screen·Senior

Senior
May 2026

Summary

Uber DS interview that was basically a full A/B testing case study crammed into one session. The question was dense, covering everything from hypothesis setup to alpha spending to CUPED variance reduction. Felt more like a take-home exam than a conversation.

Questions Asked (8)

Q1

For a primary metric defined as trip completion rate per request, state precise null and alternative hypotheses. Should the test be one-sided or two-sided, and why?

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

I went two-sided initially out of habit, then second-guessed myself mid-answer.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clearly defining the metric and the population, then state the null hypothesis as no difference in trip completion rate between control and treatment, and the alternative as a difference. Discuss the choice between one-sided and two-sided tests based on the business context and potential for both positive and negative effects, ultimately recommending a two-sided test for robustness.

Pro tip: At Uber, always consider the two-sided test as the default because it guards against unexpected negative impacts, which can be costly in a marketplace with network effects. However, if the treatment is a strict improvement with no risk of harm, a one-sided test can be justified to increase power.

1. Define the metric and population

Clarify that the metric is trip completion rate per request, calculated as the number of completed trips divided by the number of requests, and specify the population (e.g., all riders in the experiment).

2. State the null hypothesis

Null hypothesis (H0): The trip completion rate per request is equal between the control and treatment groups (p_control = p_treatment).

3. State the alternative hypothesis

Alternative hypothesis (H1): The trip completion rate per request is not equal between the control and treatment groups (p_control ≠ p_treatment) for a two-sided test, or specify direction for a one-sided test (e.g., p_treatment > p_control).

4. Choose between one-sided and two-sided

Consider the business context: if the treatment could have both positive and negative effects, use a two-sided test. If only an improvement is possible and a decrease is not a concern, a one-sided test may be appropriate. At Uber, due to network effects and potential unintended consequences, a two-sided test is generally recommended.

5. Justify the choice

Explain that a two-sided test is more conservative and guards against detecting effects in either direction, which is important for maintaining marketplace balance and rider/driver experience.

Key Points to Mention

  • Definition of trip completion rate per request as a proportion.
  • Null hypothesis: no difference in completion rates between groups.
  • Alternative hypothesis: difference in completion rates (two-sided) or directional (one-sided).
  • Two-sided test is preferred for detecting both positive and negative effects.
  • One-sided test increases power but risks missing negative impacts.
  • Business context: Uber's marketplace dynamics and network effects make two-sided testing safer.

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

Q2

Using the provided request and completion counts for both control and treatment groups, compute the z-test statistic, p-value, and a 95% confidence interval for the lift in completion rate. Show your work.

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

This is the core calculation question.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clearly defining the null and alternative hypotheses, then compute the sample proportions for both groups. Use the pooled proportion to calculate the standard error for the z-test, and the unpooled standard error for the confidence interval. Finally, compute the z-statistic, p-value, and 95% confidence interval, and interpret the results in the context of Uber's experiment.

Pro tip: Always state your assumptions (e.g., independence, large sample size) and check them. Also, clarify whether the test is one-sided or two-sided based on the business question, as this affects the p-value interpretation.

1. Define hypotheses and parameters

State the null hypothesis (no difference in completion rates) and alternative hypothesis (difference or lift). Define p_control and p_treatment as the true completion rates.

2. Compute sample proportions and standard errors

Calculate the sample completion rates for control and treatment. Compute the pooled proportion for the z-test and the unpooled standard error for the confidence interval.

3. Calculate z-statistic and p-value

Use the pooled standard error to compute the z-statistic. Find the p-value using the standard normal distribution, considering the appropriate alternative hypothesis.

4. Compute 95% confidence interval for the lift

Use the unpooled standard error to construct a 95% confidence interval for the difference in proportions (treatment - control).

5. Interpret results

Explain the practical significance of the lift, the p-value, and the confidence interval in the context of Uber's experiment. Discuss any limitations or next steps.

Key Points to Mention

  • Null and alternative hypotheses (one-sided vs two-sided)
  • Sample proportions and difference in proportions
  • Pooled vs unpooled standard error
  • Z-statistic formula and calculation
  • P-value interpretation and significance level
  • Confidence interval interpretation and practical significance

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

Q3

For the wait time guardrail metric, which test should you use and why? Compute the 95% CI for the mean difference and state your distributional assumptions.

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

Welch's t-test over pooled because the SDs differ across groups (3.2 vs 3.4) and you shouldn't assume equal variance without a reason to.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying that wait time is a guardrail metric, so the goal is to detect harm, not just significance. Recommend a two-sample t-test (Welch's) for comparing means, but also consider a non-parametric test if the distribution is heavily skewed. Then compute the 95% CI for the mean difference using the appropriate standard error and state assumptions like independence, normality (or large sample), and equal variances (if using Student's t-test).

Pro tip: Mention that guardrail metrics often require one-sided tests or non-inferiority margins to avoid false alarms, and that you'd check for outliers and consider bootstrapping if normality is questionable.

1. Clarify the metric and goal

Confirm that wait time is a continuous guardrail metric where we want to ensure the treatment doesn't significantly increase wait times. Define the null and alternative hypotheses (e.g., H0: μ_treatment - μ_control = 0 vs. H1: μ_treatment - μ_control > 0 for harm detection).

2. Choose the appropriate test

Select Welch's t-test because it does not assume equal variances and is robust to unequal sample sizes. If the data is highly skewed or has outliers, consider a non-parametric test like Mann-Whitney U or a bootstrap approach.

3. State distributional assumptions

Assumptions: independent observations, approximately normal sampling distribution of the mean difference (or large sample sizes due to CLT), and for Welch's t-test, no assumption of equal variances. Mention that if sample sizes are small and data non-normal, use bootstrap.

4. Compute the 95% CI for mean difference

Calculate the difference in sample means, the standard error using Welch's formula, and the degrees of freedom. Then use the t-distribution to find the critical value and construct the CI: (x̄1 - x̄2) ± t* * SE.

5. Interpret and act

Interpret the CI: if it includes zero, no significant difference; if it lies entirely above zero, the treatment may harm wait times. Consider practical significance and guardrail thresholds.

Key Points to Mention

  • Welch's t-test for unequal variances
  • Assumptions: independence, normality (or large n), no equal variance assumption
  • 95% CI formula: difference in means ± t* * SE
  • Alternative non-parametric tests (Mann-Whitney U) or bootstrap for non-normal data
  • One-sided vs two-sided testing for guardrail metrics
  • Practical significance and guardrail thresholds

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

Q4

Apply Holm-Bonferroni correction across all three outcomes (primary metric, cancel rate guardrail, wait time guardrail) at a familywise alpha of 0.05. Which effects survive?

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

I knew the procedure but blanked on the ordering step for a second.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

First, clarify the raw p-values for each of the three outcomes and confirm they are from independent tests. Then apply Holm-Bonferroni by ordering the p-values from smallest to largest, comparing each to alpha/(m - rank + 1), and stopping at the first non-significant result. Finally, report which effects survive the correction and discuss the implications for decision-making.

Pro tip: In practice, guardrail metrics often have lower power or higher variance, so a non-significant guardrail after correction doesn't necessarily mean it's safe—consider practical significance and confidence intervals alongside p-values. Also, be prepared to explain why Holm-Bonferroni is preferred over Bonferroni for its greater power while still controlling FWER.

1. List and order the p-values

Obtain the raw p-values for the primary metric, cancel rate guardrail, and wait time guardrail. Order them from smallest to largest: p(1) ≤ p(2) ≤ p(3).

2. Determine the Holm-adjusted thresholds

For each rank i (1 to 3), compute the threshold α/(m - i + 1) where m=3 and α=0.05. So thresholds are 0.05/3 ≈ 0.0167, 0.05/2 = 0.025, and 0.05/1 = 0.05.

3. Compare and stop at first non-significance

Compare p(1) to 0.0167; if significant, compare p(2) to 0.025; if significant, compare p(3) to 0.05. Stop at the first p-value that exceeds its threshold; all subsequent hypotheses are not rejected.

4. Interpret surviving effects

Identify which outcomes have p-values below their respective thresholds. These are the effects that survive Holm-Bonferroni correction at FWER = 0.05.

5. Discuss practical implications

Explain what the surviving effects mean for the experiment: e.g., if the primary metric survives but a guardrail does not, consider whether the guardrail is practically violated despite lack of statistical significance.

Key Points to Mention

  • Holm-Bonferroni is a step-down procedure that controls the familywise error rate (FWER) while being uniformly more powerful than Bonferroni.
  • The correction accounts for multiple comparisons, reducing the chance of false positives when testing multiple metrics.
  • The order of p-values matters: the smallest p-value is compared to the strictest threshold, and thresholds become less strict for larger p-values.
  • If a p-value fails to meet its threshold, all larger p-values are also considered non-significant (the step-down stops).
  • Guardrail metrics are typically one-sided tests (non-inferiority) and may require different handling, but Holm-Bonferroni can still be applied if p-values are computed appropriately.
  • Always report effect sizes and confidence intervals alongside p-values to assess practical significance, especially for guardrails.

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

Q5

In plain language, what does the p-value you computed actually mean, and what does it not mean?

A/B Testing & Experimentation
Author's notes

Straightforward but easy to over-explain.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by defining the p-value in plain language: it's the probability of observing data as extreme or more extreme than what you saw, assuming the null hypothesis is true. Then clearly state what it does not mean—it is not the probability that the null hypothesis is true, nor does it measure effect size or practical significance. Finally, connect this to A/B testing at Uber by emphasizing that decisions should combine p-values with effect sizes, confidence intervals, and business context.

Pro tip: Mention that with Uber's massive sample sizes, tiny and practically irrelevant effects can produce very small p-values, so always pair statistical significance with practical significance and guard against peeking or multiple testing.

1. Define the p-value plainly

State that the p-value is the probability of seeing an effect at least as extreme as observed, assuming the null hypothesis (no effect) is true. Avoid jargon like 'conditional probability' unless you explain it.

2. Clarify what it does NOT mean

Emphasize that it is not the probability that the null hypothesis is true, nor the probability that results are due to chance, nor a measure of effect size or importance.

3. Connect to A/B testing context

Explain that in A/B tests, a small p-value suggests the observed difference is unlikely under the null, but it doesn't tell you if the difference is large enough to matter for the business.

4. Advocate for complementary metrics

Recommend using effect sizes, confidence intervals, and business metrics (e.g., revenue, retention) alongside p-values to make informed decisions.

5. Highlight pitfalls at scale

Note that large samples can yield tiny p-values for trivial effects, and that p-hacking, peeking, and multiple comparisons inflate false positives.

Key Points to Mention

  • P-value is the probability of observing data as extreme or more extreme than actual, given the null hypothesis is true.
  • It is not the probability that the null hypothesis is true or that the result is due to chance.
  • It does not measure the size or practical importance of an effect.
  • In A/B testing, combine p-values with effect sizes, confidence intervals, and business impact.
  • Large sample sizes can make trivial effects statistically significant, so practical significance matters.
  • Avoid p-hacking, peeking, and multiple testing issues by pre-registering analyses and using corrections.

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

Q6

The experiment had five unplanned interim looks at the data with no pre-registered alpha spending. How does this affect your significance conclusion, and how would you apply a Pocock or O'Brien-Fleming approach to address it?

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

This one tripped me up more than I expected.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Acknowledge that the unplanned interim looks inflate the Type I error rate, so the original p-value is no longer valid. Explain how to apply Pocock or O'Brien-Fleming alpha spending to correct the significance threshold, and discuss the trade-offs between the two approaches. Conclude with a recommendation on how to interpret the results and what to do next.

Pro tip: Mention that Pocock is better for detecting early large effects, while O'Brien-Fleming is more conservative early and preserves power for the final analysis—choose based on the study's goals and expected effect timing.

1. Acknowledge the issue

State that unplanned interim looks without alpha spending inflate the family-wise error rate (FWER), making the naive p-value invalid.

2. Explain the impact

Quantify how the significance conclusion changes: the effective alpha is higher than 0.05, so the observed effect may not be statistically significant after correction.

3. Apply alpha spending

Describe how to use Pocock or O'Brien-Fleming boundaries to adjust the significance threshold at each look, and compute the adjusted p-value or critical value.

4. Compare approaches

Contrast Pocock (constant boundaries, more power for early effects) and O'Brien-Fleming (conservative early, more power at the end) in terms of Type I error control and power.

5. Recommend and conclude

Suggest which method fits the context (e.g., O'Brien-Fleming for unknown timing) and state the corrected conclusion, possibly recommending a follow-up experiment.

Key Points to Mention

  • Family-wise error rate (FWER) inflation due to multiple testing
  • Alpha spending functions and how they control Type I error
  • Pocock boundaries: constant alpha at each look, more power for early stopping
  • O'Brien-Fleming boundaries: conservative early, more power at final analysis
  • Trade-offs between Pocock and O'Brien-Fleming in terms of power and timing
  • Adjusted p-value or critical value calculation and interpretation

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

Q7

If the pre-experiment completion rate has a correlation of 0.40 with the in-experiment outcome, estimate the variance reduction from CUPED and explain how it would affect required sample size or interpretation of results.

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

CUPED variance reduction is 1 minus r-squared, so with r=0.40 you get roughly 16% variance reduction.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by recalling the CUPED variance reduction formula: the reduction in variance is equal to the square of the correlation between the pre-experiment covariate and the outcome. Then compute the variance reduction as 0.40^2 = 0.16, meaning a 16% reduction in variance. Finally, explain that this reduction translates to a proportional decrease in required sample size (since sample size is proportional to variance) and discuss how it affects the interpretation of results, such as increased power or narrower confidence intervals.

Pro tip: Emphasize that CUPED uses pre-experiment data to reduce variance without introducing bias, and that the variance reduction is directly proportional to the square of the correlation. Also, mention that while the sample size reduction is substantial, it's important to ensure the pre-experiment covariate is unaffected by the treatment and is measured before the experiment.

1. Recall CUPED variance reduction formula

State that the proportional variance reduction from CUPED is given by ρ^2, where ρ is the correlation between the pre-experiment covariate and the outcome metric.

2. Compute the variance reduction

Plug in ρ = 0.40 to get ρ^2 = 0.16, meaning a 16% reduction in variance (or equivalently, the variance is multiplied by 0.84).

3. Relate variance reduction to sample size

Explain that required sample size is proportional to variance, so a 16% reduction in variance leads to approximately a 16% reduction in required sample size (or equivalently, the same sample size yields higher power).

4. Discuss impact on interpretation

Mention that with CUPED, confidence intervals become narrower and p-values smaller for the same effect size, potentially leading to earlier stopping or detection of smaller effects.

5. Note assumptions and caveats

Highlight that CUPED requires the pre-experiment covariate to be uncorrelated with treatment assignment and measured before the experiment; otherwise, bias may be introduced.

Key Points to Mention

  • CUPED variance reduction formula: reduction = ρ^2
  • Calculation: 0.40^2 = 0.16 (16% variance reduction)
  • Sample size reduction: 16% fewer users needed for same power
  • Increased sensitivity: ability to detect smaller effect sizes
  • Narrower confidence intervals and more precise estimates
  • Assumption: pre-experiment covariate must be independent of treatment and measured pre-treatment

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

Q8

Given all the analysis above, including the guardrail results, should you ship the feature, iterate, or stop? Defend your recommendation.

A/B Testing & ExperimentationProduct Sense & IdeationTechnical Trade-offs
Author's notes

The cancel rate going up is the real tension here.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by restating the decision options and the key evidence from the analysis, including guardrail results. Then weigh the trade-offs between statistical significance, practical impact, and business risk to make a clear recommendation. Defend it by acknowledging limitations and outlining next steps or monitoring plans.

Pro tip: Show that you understand the difference between statistical significance and practical significance, and always tie your recommendation back to Uber's core metrics like rider/driver experience, marketplace efficiency, and long-term growth.

1. Summarize the evidence

Briefly recap the primary metric results, guardrail metrics, and any segment-level findings from the experiment. Highlight whether the results are statistically significant and practically meaningful.

2. Evaluate trade-offs

Discuss the balance between potential gains in the primary metric and any negative impacts on guardrails (e.g., rider cancellations, driver earnings, safety). Consider short-term vs. long-term effects and scalability.

3. Make a recommendation

Choose one of the three options—ship, iterate, or stop—based on the evidence and trade-offs. Be decisive and align your choice with Uber's strategic priorities.

4. Defend with reasoning

Justify your recommendation by referencing specific data points, business context, and potential risks. Address counterarguments and explain why your choice is optimal.

5. Outline next steps

If shipping, describe monitoring and rollback plans. If iterating, specify what changes to test next. If stopping, suggest alternative approaches or learnings to apply.

Key Points to Mention

  • Statistical significance vs. practical significance: ensure the effect size is meaningful for Uber's scale.
  • Guardrail metrics: any degradation in rider/driver satisfaction, safety, or marketplace health is a red flag.
  • Segment analysis: check if the effect varies across cities, user types, or time periods to avoid Simpson's paradox.
  • Business impact: estimate the potential revenue or cost implications and alignment with Uber's strategic goals.
  • Risk mitigation: propose phased rollout, holdback groups, or additional experiments to de-risk the decision.
  • Long-term vs. short-term: consider novelty effects and whether the impact will persist.

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