← Weride Interview Insights

Weride·Data Scientist·Technical Phone Screen·Senior

SeniorPrefer not to say
Jul 2026

Summary

WeRide data scientist interview, technical phone screen that went deep fast. Two big questions, both stat-heavy, and I felt underprepared on the sampling side.

Questions Asked (2)

Q1

You have two datasets from different sources (e.g. two system versions, cities, or time periods). How do you test whether they come from the same underlying distribution? Walk through continuous vs categorical variables, parametric vs nonparametric tests and their assumptions, visual diagnostics, effect sizes, statistical vs practical significance, multiple testing, and what changes when sample sizes are very unequal.

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

This one took a while to unpack.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying the goal: are you testing for any distributional difference or specific shifts? Then systematically address continuous and categorical variables with appropriate visual and statistical methods, considering assumptions, effect sizes, and multiple testing. Finally, discuss how unequal sample sizes affect test choice and interpretation, emphasizing practical significance over p-values alone.

Pro tip: Always pair statistical tests with effect size measures and visualizations; a tiny p-value with a trivial effect size is common with large samples and can mislead stakeholders. Also, consider bootstrapping or permutation tests when assumptions are dubious, especially with unequal sample sizes.

1. Clarify objective and data types

Determine whether the goal is to detect any difference or a specific shift, and separate variables into continuous and categorical. This guides the choice of tests and diagnostics.

2. Visual diagnostics and exploratory analysis

Use histograms, density plots, boxplots, Q-Q plots for continuous variables; bar charts and mosaic plots for categorical. These reveal distribution shape, outliers, and potential differences before formal testing.

3. Select and apply statistical tests

For continuous: if normality and equal variance hold, use t-test (mean) or F-test (variance); otherwise use nonparametric tests like Mann-Whitney U (median/shift), Kolmogorov-Smirnov (any difference), or Anderson-Darling. For categorical: use chi-square or Fisher's exact test. Adjust for multiple comparisons (e.g., Bonferroni, FDR).

4. Quantify effect sizes and practical significance

Compute Cohen's d, Cliff's delta, Cramér's V, or overlap measures. Interpret in context: a statistically significant difference may be practically irrelevant if effect size is small.

5. Address unequal sample sizes and robustness

With very unequal sizes, parametric tests may be sensitive to variance differences; use Welch's t-test or nonparametric alternatives. Consider permutation tests or bootstrapping to assess differences in distributions without strict assumptions.

Key Points to Mention

  • Parametric vs nonparametric tests and their assumptions (normality, homogeneity of variance, independence).
  • Visual diagnostics: Q-Q plots, density plots, boxplots, and empirical CDFs.
  • Effect size measures: Cohen's d, Cliff's delta, Cramér's V, and overlap coefficient.
  • Multiple testing corrections: Bonferroni, Benjamini-Hochberg FDR.
  • Statistical vs practical significance: p-values vs effect sizes and domain context.
  • Unequal sample sizes: use of Welch's t-test, permutation tests, and caution with KS test sensitivity.

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

Q2

What sampling methods do you know and when would you use each? Cover simple random, stratified, cluster, systematic, and weighted or importance sampling at minimum, and say which you'd prefer for product evaluation.

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

Blanked a little on importance sampling, which is embarrassing given the autonomous driving context where it's actually super relevant for rare event evaluation.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by briefly defining each sampling method and its typical use case, then explain when each is appropriate based on data characteristics and goals. For product evaluation, recommend stratified or weighted sampling to ensure representative and unbiased results, and justify your choice with practical considerations like cost, accuracy, and feasibility.

Pro tip: Emphasize that the choice of sampling method depends on the specific product metrics and potential biases; for A/B testing, stratified sampling by key user segments often yields more precise treatment effect estimates than simple random sampling.

1. Define each sampling method

Clearly explain simple random, stratified, cluster, systematic, and weighted/importance sampling, including how each is implemented.

2. Explain when to use each method

Describe scenarios where each method is most appropriate, considering factors like population homogeneity, cost, and need for representation.

3. Compare trade-offs

Discuss pros and cons such as bias, variance, cost, and practicality for each method in the context of data science.

4. Recommend for product evaluation

State your preferred method(s) for product evaluation, typically stratified or weighted sampling, and justify why they are suitable for A/B testing and experimentation.

5. Connect to business impact

Tie your recommendation to how it improves decision-making, reduces bias, and aligns with company goals like accurate measurement of product changes.

Key Points to Mention

  • Simple random sampling: unbiased but may miss small subgroups; use when population is homogeneous and complete list is available.
  • Stratified sampling: ensures representation of key segments; use when subgroups differ and you want precise estimates per group.
  • Cluster sampling: cost-effective for geographically dispersed populations; use when natural groups exist and listing individuals is impractical.
  • Systematic sampling: easy to implement with a random start; use when population is ordered and you want a spread across the list.
  • Weighted/importance sampling: corrects for underrepresentation or focuses on rare events; use when certain groups are oversampled or when estimating rare outcomes.
  • For product evaluation, stratified or weighted sampling is preferred to ensure balanced A/B test groups and accurate treatment effect estimation.

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