← Coinbase Interview Insights

Coinbase·Data Scientist·Technical Phone Screen·Intermediate

Intermediate
May 2026

Summary

Probability and stats question for a data scientist role at Coinbase. Pretty textbook stuff but the conditional probability part at the end tripped me up more than I expected.

Questions Asked (3)

Q1

If n users each independently link a crypto wallet with probability p, what is the expected number of users who link their wallet?

Product Analytics & MetricsAlgorithms & Data Structures
Author's notes

This one was fine.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Recognize that each user's wallet linking is an independent Bernoulli trial with success probability p. The total number of linked users is the sum of n independent Bernoulli variables, so the expected value is n * p. State the result clearly and explain the reasoning.

Pro tip: Mention that this is a binomial distribution and that the expectation is linear, so you don't need independence for the expectation to hold—though independence is given here. This shows you understand the underlying assumptions and can generalize.

1. Define the random variable

Let X_i be an indicator variable that equals 1 if user i links their wallet, and 0 otherwise. Then X_i ~ Bernoulli(p).

2. Express total linked users

The total number of linked users is S = X_1 + X_2 + ... + X_n.

3. Apply linearity of expectation

Since expectation is linear, E[S] = E[X_1] + E[X_2] + ... + E[X_n] = n * p.

4. State the result and distribution

Thus, the expected number is n * p. Also note that S follows a Binomial(n, p) distribution, so the variance is n * p * (1-p).

Key Points to Mention

  • Independence of user actions
  • Bernoulli trials and indicator variables
  • Linearity of expectation
  • Binomial distribution (n, p)
  • Expected value formula: n * p
  • Variance: n * p * (1-p) (optional but shows depth)

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

Q2

What is the probability that at least one of the n users links their wallet?

Product Analytics & MetricsAlgorithms & Data Structures
Author's notes

Complement rule, so 1 minus the probability nobody links, which is 1 minus (1-p)^n.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Clarify the assumptions about independence and the probability of a single user linking their wallet. Then model the number of users who link as a binomial random variable and compute the probability of at least one success as 1 minus the probability of zero successes. If the per-user probability is not given, denote it as p and express the answer in terms of p and n.

Pro tip: Always state your assumptions explicitly and consider edge cases like p=0 or p=1; also mention that in real product analytics, user behaviors are often not independent, so you might need to account for correlation or use historical data to estimate p.

1. Clarify the problem and assumptions

Ask clarifying questions to confirm whether each user links independently and whether the probability of linking is the same for all users. If not provided, denote the probability as p.

2. Define the random variable

Let X be the number of users who link their wallet out of n. Under the assumptions, X follows a Binomial(n, p) distribution.

3. Compute the probability of at least one

Use the complement rule: P(X ≥ 1) = 1 - P(X = 0) = 1 - (1 - p)^n.

4. Discuss edge cases and extensions

Consider what happens if p is very small or n is large, and mention that if p is unknown, it must be estimated from data. Also note that independence may not hold in practice.

Key Points to Mention

  • Independence assumption and identical distribution of user behavior
  • Binomial distribution and its parameters (n, p)
  • Complement rule for 'at least one' probability
  • Formula: 1 - (1 - p)^n
  • Edge cases: p=0, p=1, large n, small p
  • Real-world considerations: correlation among users, varying p, and estimating p from historical data

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

Q3

Given that at least one of two specific users A or B links a wallet, what is the probability that both of them do?

Product Analytics & MetricsA/B Testing & ExperimentationAlgorithms & Data Structures
Author's notes

This is where I slowed down.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Clarify the assumptions about the joint distribution of wallet linking between users A and B, then apply conditional probability. If independence is assumed, compute P(A and B) / P(A or B) using given or estimated marginal probabilities; otherwise, discuss how correlation or additional data would be needed.

Pro tip: In real-world product analytics, user behaviors are rarely independent; always check for correlation (e.g., via historical data or cohort analysis) before assuming independence. Mentioning this shows you understand practical data nuances.

1. Define the events and given condition

Let A be the event that user A links a wallet, B for user B. The condition is that at least one links, i.e., A ∪ B. We want P(A ∩ B | A ∪ B).

2. Apply conditional probability formula

Use P(A ∩ B | A ∪ B) = P(A ∩ B) / P(A ∪ B). Expand P(A ∪ B) = P(A) + P(B) - P(A ∩ B).

3. Assess independence assumption

If A and B are independent, P(A ∩ B) = P(A)P(B). Then the probability simplifies to P(A)P(B) / (P(A)+P(B)-P(A)P(B)). If not independent, you need the joint probability or correlation.

4. Estimate or obtain probabilities

Use historical data to estimate P(A), P(B), and possibly P(A ∩ B). If only marginal probabilities are available, state the independence assumption explicitly and compute.

5. Interpret and discuss limitations

Present the computed probability and discuss how violations of independence (e.g., due to referral, shared device, or temporal factors) could affect the result.

Key Points to Mention

  • Conditional probability formula: P(A ∩ B | A ∪ B) = P(A ∩ B) / P(A ∪ B)
  • Inclusion-exclusion principle for P(A ∪ B)
  • Independence assumption and its implications
  • Real-world dependence between users (e.g., referrals, shared networks)
  • Need for joint probability data when independence fails
  • Bayesian or empirical estimation from historical data

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