← Lyft Interview Insights

Lyft·Data Scientist·Technical Phone Screen·Intermediate

Intermediate
May 2026

Summary

Interviewed for a Data Scientist role at Lyft. One probability question that sounds straightforward until you realize you need to set up the conditional correctly.

Questions Asked (1)

Q1

A rider takes Lyft home from work. What is the probability they also took Lyft to work that morning?

Product Analytics & Metrics
Author's notes

Classic conditional probability setup.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Clarify that this is a conditional probability question asking for P(took Lyft to work | took Lyft home). Recognize that the answer depends on assumptions about rider behavior, so state a reasonable assumption (e.g., morning and evening choices are independent) and compute the probability using Bayes' theorem or a simple 2x2 table. If independence is assumed, the probability equals the overall share of commuters who take Lyft to work.

Pro tip: Don't jump to a number—first ask clarifying questions about the data available and the rider population. Then explicitly state your assumptions (e.g., independence) and show how the answer changes if those assumptions are relaxed, demonstrating analytical rigor.

1. Clarify the question

Confirm that the question asks for the conditional probability that a rider took Lyft to work given they took Lyft home. Ask whether we have historical data on commute patterns or if we should assume a model.

2. Define variables and assumptions

Let A = took Lyft to work, B = took Lyft home. We want P(A|B). Assume independence between morning and evening choices unless told otherwise, meaning P(A|B) = P(A).

3. Estimate P(A)

Estimate the overall probability that a commuter takes Lyft to work. This could be based on Lyft's market share for morning commutes or internal data. For example, if 20% of commuters use Lyft in the morning, then P(A) = 0.2.

4. Compute the conditional probability

Under independence, P(A|B) = P(A) = 0.2 (or whatever estimate). If independence is unrealistic, use Bayes' theorem: P(A|B) = P(B|A) * P(A) / P(B), and discuss how to estimate P(B|A) and P(B).

5. Discuss limitations and extensions

Acknowledge that morning and evening choices are likely correlated (e.g., due to weather, car availability, or habit). Suggest using historical data to compute the actual conditional probability or building a model that accounts for correlation.

Key Points to Mention

  • Conditional probability and Bayes' theorem
  • Independence assumption and its limitations
  • Base rate (prior probability) of taking Lyft to work
  • Data sources: internal Lyft data, surveys, or market research
  • Correlation between morning and evening commute modes
  • Business implications: rider retention, subscription models, or incentives

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