← Jane Street Interview Insights

Jane Street·Data Scientist·Technical Phone Screen·Intermediate

IntermediatePrefer not to say
Jun 2026

Summary

Jane Street data scientist interview, all probability puzzles, no coding, no case studies. The interviewer just handed me five problems and expected me to think out loud the whole time. Felt more like a math exam than a job interview, which I was not fully prepared for.

Questions Asked (5)

Q1

You eat pizza on Saturday with probability 0.4 and on Sunday with probability 0.3. No information is given about how the two days relate. What is the range of possible values for the probability of eating pizza at least once over the weekend? Describe a dependence structure that achieves each extreme.

Algorithms & Data Structures
Author's notes

I jumped straight to the independence formula and got a number, then the interviewer asked me to slow down and think about what the actual bounds were.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Use the inclusion-exclusion principle to express the probability of at least one pizza day as P(Sat) + P(Sun) - P(Sat and Sun). Recognize that the only unknown is the joint probability, which is constrained by the marginal probabilities. Then find the minimum and maximum possible values of the joint probability and construct dependence structures that achieve them.

Pro tip: Emphasize that the range is determined by the Fréchet–Hoeffding bounds, and explicitly describe the dependence structures (e.g., mutually exclusive events for the minimum, and one event implying the other for the maximum). This shows you understand both the math and the practical implications of dependence.

1. Define events and target probability

Let A be the event of eating pizza on Saturday, B on Sunday. We want P(A ∪ B). Use inclusion-exclusion: P(A ∪ B) = P(A) + P(B) - P(A ∩ B) = 0.4 + 0.3 - P(A ∩ B) = 0.7 - P(A ∩ B).

2. Determine bounds on the joint probability

The joint probability P(A ∩ B) must satisfy max(0, P(A)+P(B)-1) ≤ P(A ∩ B) ≤ min(P(A), P(B)). Here, max(0, 0.7-1) = 0, and min(0.4, 0.3) = 0.3. So 0 ≤ P(A ∩ B) ≤ 0.3.

3. Compute the range for P(A ∪ B)

Substitute the bounds into the inclusion-exclusion formula: minimum P(A ∪ B) = 0.7 - 0.3 = 0.4; maximum P(A ∪ B) = 0.7 - 0 = 0.7. Thus the range is [0.4, 0.7].

4. Describe dependence structures for each extreme

For the minimum (0.4), make A and B mutually exclusive (P(A ∩ B)=0). For the maximum (0.7), make B a subset of A (P(A ∩ B)=0.3), so that Sunday pizza always implies Saturday pizza.

Key Points to Mention

  • Inclusion-exclusion principle for union of events.
  • Fréchet–Hoeffding bounds for joint probability given marginals.
  • Mutually exclusive events achieve the minimum union probability.
  • One event being a subset of the other achieves the maximum union probability.
  • The range is [0.4, 0.7].
  • Dependence structure affects the probability of at least one event.

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

Q2

You flip 250 gold coins and 250 silver coins. Gold heads score 3 points, silver heads score 1 point, tails score 0. Given that the total score is exactly 900, which combination of gold heads and silver heads is most likely?

Algorithms & Data Structures
Author's notes

The constraint 3g + s = 900 pins s once you pick g, so you're really maximizing a product of two binomial coefficients.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Model the total score as a sum of independent random variables: gold heads ~ Binomial(250, 0.5) and silver heads ~ Binomial(250, 0.5). The total score S = 3G + H. Given S = 900, find the conditional distribution of G (or H) and identify the mode. Use the fact that G and H are independent and compute P(G=g, H=h | S=900) ∝ P(G=g)P(H=h) with 3g + h = 900.

Pro tip: Leverage the symmetry and the fact that the conditional distribution is approximately normal; the mode occurs where the quadratic exponent is maximized, which can be found by solving a simple equation. Also, note that the most likely combination is not necessarily the one that maximizes the individual probabilities without conditioning.

1. Define random variables

Let G be the number of gold heads and H be the number of silver heads. Then G ~ Binomial(250, 0.5) and H ~ Binomial(250, 0.5), independent. The total score is S = 3G + H.

2. Set up conditional probability

We want to find (g, h) that maximizes P(G=g, H=h | S=900). Since S=900 implies h = 900 - 3g, we need to maximize P(G=g) * P(H=900-3g) over integer g such that 0 ≤ g ≤ 250 and 0 ≤ 900-3g ≤ 250.

3. Approximate with normal distributions

Use normal approximations: G ≈ N(125, 62.5) and H ≈ N(125, 62.5). Then the log-likelihood is proportional to -(g-125)^2/(2*62.5) - (h-125)^2/(2*62.5) with h=900-3g. Maximize this quadratic.

4. Solve for the mode

Differentiate the quadratic with respect to g and set to zero: -(g-125)/62.5 - 3*(h-125)/62.5 * (-3?) Wait, careful: h = 900 - 3g, so dh/dg = -3. The derivative of the log-likelihood is -(g-125)/62.5 + 3*(h-125)/62.5 = 0. Solve for g, then h = 900 - 3g.

5. Check integer and boundaries

The solution may not be integer; check nearby integers. Also ensure g and h are within [0,250]. The most likely combination is the one that maximizes the product of binomial probabilities.

Key Points to Mention

  • Independence of gold and silver coin flips.
  • Binomial distribution for number of heads.
  • Conditional probability and the constraint 3G + H = 900.
  • Normal approximation to binomial for large n.
  • Maximizing the product of probabilities (or log-probabilities) subject to the constraint.
  • The mode of the conditional distribution is found by solving a quadratic equation.

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

Q3

An ant starts at one vertex of a unit cube and moves each second to a uniformly random adjacent vertex. What is the expected number of steps to reach the vertex diagonally opposite the starting point?

Algorithms & Data Structures
Author's notes

Classic random walk on a cube.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Model the ant's movement as a Markov chain on the cube's vertices, grouping states by distance from the start (0, 1, 2, 3). Set up equations for the expected hitting time from each distance and solve them to find the expected steps from the start to the opposite vertex.

Pro tip: Emphasize that the cube's symmetry allows you to collapse the state space to just four states, which is a powerful technique for simplifying complex random walks. Also, verify your answer by checking that the expected time from the opposite vertex is zero and that the equations are consistent.

1. Define states by distance

Label the starting vertex as state 0, vertices adjacent to start as state 1, vertices adjacent to the opposite as state 2, and the opposite vertex as state 3. Note that the ant moves to a uniformly random adjacent vertex each second.

2. Determine transition probabilities

From state 0, the ant always moves to state 1. From state 1, it moves to state 0 with probability 1/3 and to state 2 with probability 2/3. From state 2, it moves to state 1 with probability 2/3 and to state 3 with probability 1/3. State 3 is absorbing.

3. Set up expected hitting time equations

Let E_i be the expected number of steps to reach state 3 from state i. Write equations: E_3 = 0, E_2 = 1 + (2/3)E_1 + (1/3)E_3, E_1 = 1 + (1/3)E_0 + (2/3)E_2, E_0 = 1 + E_1.

4. Solve the system of equations

Substitute E_3 = 0 into E_2, then solve the linear system for E_0. The solution yields E_0 = 10 steps.

5. Verify and interpret

Check that the solution satisfies all equations and makes intuitive sense: the ant must pass through intermediate states, and the expected time is finite. The answer is 10 steps.

Key Points to Mention

  • Markov chain modeling and state space reduction using symmetry
  • Expected hitting time and first-step analysis
  • Transition probabilities based on cube connectivity
  • Solving linear equations for expected values
  • Verification of the solution by substitution or simulation
  • Interpretation of the result in the context of the problem

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

Q4

You have N shoelaces in a pile with 2N free ends. Repeatedly pick two free ends at random and tie them together until no free ends remain. What is the expected number of closed loops formed?

Algorithms & Data Structures
Author's notes

Linearity of expectation saves you here.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Use linearity of expectation by defining indicator variables for each potential loop closure. Recognize that the process is equivalent to a random pairing of the 2N ends, and compute the probability that a given set of ends forms a loop, then sum over all possible loops.

Pro tip: Start with small N (e.g., N=1,2,3) to identify the pattern and verify your formula. This demonstrates systematic thinking and helps catch mistakes early.

1. Understand the process

Clarify that we randomly pair all 2N ends, and a loop is formed when a sequence of ties connects ends of the same lace or previously connected components.

2. Define indicator variables

For each possible loop, define an indicator variable that is 1 if that loop is formed. The total number of loops is the sum of these indicators.

3. Compute probability of a specific loop

Calculate the probability that a given set of ends forms a loop. For a loop involving k laces, the probability is 1/(2N-1) * 1/(2N-3) * ... * 1/(2k-1) times the number of ways to order the laces.

4. Sum over all possible loops

Sum the probabilities over all possible loops. Use combinatorial counting to account for the number of loops of each size and simplify the sum.

5. Simplify to closed form

Evaluate the sum to obtain the expected number of loops, which is the Nth harmonic number H_N = 1 + 1/2 + ... + 1/N.

Key Points to Mention

  • Linearity of expectation allows summing probabilities without worrying about dependencies.
  • The process is equivalent to a uniformly random perfect matching of the 2N ends.
  • The probability that a specific set of k laces forms a loop is 2^{k-1} (k-1)! / (2k-1)!!.
  • The number of ways to choose k laces from N is C(N, k).
  • The sum simplifies to the harmonic number H_N.
  • For large N, the expected number grows logarithmically.

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

Q5

Two equal-volume buckets, one pure white and one pure red. Pour a quarter bucket of red into white, mix, then pour the same volume of the mixture back into the red bucket and mix again. Which bucket's color changed more?

Algorithms & Data Structures
Author's notes

Neither.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Recognize that this is a conservation problem: the amount of red in the white bucket equals the amount of white in the red bucket after the two transfers. Therefore, both buckets experience the same magnitude of color change, just in opposite directions. Explain this using the principle of conservation of volume and concentration, avoiding unnecessary calculations.

Pro tip: Emphasize that the answer is independent of the volumes and mixing ratios, as long as the final volumes are equal. This demonstrates strong quantitative reasoning and the ability to generalize.

1. Understand the setup

Identify the initial state: two equal-volume buckets, one pure white, one pure red. Note the sequence: transfer a quarter bucket of red to white, mix, then transfer the same volume of mixture back to red.

2. Track the transfers

Let the initial volume of each bucket be V. After the first transfer, the white bucket has V + V/4 = 5V/4 of mixture, and the red bucket has 3V/4 of pure red.

3. Analyze the second transfer

The second transfer moves V/4 of the mixture from the white bucket back to the red bucket. This mixture contains both red and white in some proportion.

4. Apply conservation

Since the final volumes are equal (V each), the amount of red in the white bucket must equal the amount of white in the red bucket. Thus, the color change is symmetric.

5. Conclude

Both buckets undergo the same magnitude of color change: the white bucket gains red, the red bucket gains white, and the amounts are equal.

Key Points to Mention

  • Conservation of volume: final volumes are equal, so the amount of red in white equals the amount of white in red.
  • The answer is independent of the initial volumes and the transfer amount, as long as the final volumes are equal.
  • The color change is symmetric: both buckets change by the same amount, just in opposite directions.
  • This is a classic problem illustrating the importance of conservation laws in quantitative reasoning.
  • Avoid getting bogged down in calculating exact concentrations; focus on the invariant.
  • The mixing ensures homogeneity, but the result holds even without perfect mixing as long as the final volumes are equal.

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