SIG (Susquehanna) Data Scientist interview questions

12 Data Scientists have shared notes from their SIG (Susquehanna) interviews, with the questions they were asked in each round. That is 19% of the 63 SIG (Susquehanna) interview notes on Screna AI.

12 Data Scientist interview notes

SIG (Susquehanna)·Data Scientist·Technical Phone Screen

Jul 2026
SIG probability question for a Data Scientist role. One problem, no fluff, just math under pressure.
  • A bakery has 2 muffins left and 5 customers still in line. Each customer independently buys a muffin with 30% probability and a croissant with 70%. What is the probability that the muffins don't run out before everyone is served?

“This one took me longer than it should have.”

View Post

SIG (Susquehanna)·Data Scientist·Technical Phone Screen

Jul 2026
SQL-heavy technical screen for a data scientist role at SIG. Six questions back to back, all analytics SQL, no behavioral fluff. The problems ranged from basic conversion funnels to some genuinely tricky window function and cross-join completeness stuff.
  • For a fixed 7-day window, compute conversion rate (delivered / created) broken down by city and courier type. An order counts as delivered if status is 'delivered' and delivered_at is not null, regardless of when it was delivered.
  • Generate a monthly volume report for the last 3 months showing orders created, orders delivered, and GMV (delivered count times a fixed rate) per courier type. Months with zero activity must still appear in the output.
  • For each user, find the first streak of 3 consecutive delivered biker orders. Any non-biker delivery breaks the streak; non-delivered orders are ignored entirely. Return the first and third order IDs and the time span between them in minutes.
  • Compute the 95th percentile delivery time in minutes for delivered orders created in the last 30 days, grouped by courier type. Also return the count per group.
  • Extend the monthly volume report to include city as a dimension, ensuring every combination of city (from both users and couriers), courier type, and month appears in the output even with zero orders.
  • Find all orders created on or before a cutoff timestamp that were not delivered within 2 hours of creation. Include orders where delivered_at is null or where it exceeds the 2-hour threshold. Return a boolean column flagging these stale orders.

“Straightforward on the surface but the 'regardless of when delivered' part tripped me up briefly.”

View Post

SIG (Susquehanna)·Data Scientist·Technical Phone Screen

Jul 2026
A probability brain teaser from SIG for a Data Scientist role. Just the one question from what I can tell, classic quant-style stuff you'd expect from a trading firm.
  • Two flowers bloom independently. The purple flower starts blooming at a uniformly random time over the next 30 days and stays in bloom for 9 days. The red flower also starts at a uniformly random time over the same 30-day window and stays in bloom for 12 days. What is the probability that they overlap in bloom at some point?

“This is a geometric probability problem.”

View Post

SIG (Susquehanna)·Data Scientist·Technical Phone Screen

Jun 2026
SIG logic puzzle interview for a Data Scientist role. One question, four variants, and the kind of problem that sounds like a riddle from a philosophy class. Pretty classic for a quant-adjacent shop.
  • On an island of knights (always truthful) and knaves (always lying), person B says to A: 'You always lie,' and then makes a second claim. There are four possible second claims: (1) A is a knave and B is a knight, (2) A is a knight and B is a knave, (3) A is a knave and B is a knave, (4) A is a knight and B is a knight. For each variant, determine the consistent identities of A and B, or state that no consistent solution exists.

“This wrecked me a little.”

View Post

SIG (Susquehanna)·Data Scientist·Technical Phone Screen

Jun 2026
A brain teaser style math problem from a SIG Data Scientist interview. Short and self-contained but the kind of thing that trips you up if you rush the setup.
  • A barn contains spiders, chickens, and cows. Spiders have 8 legs, chickens have 2, and cows have 4. The total leg count is 520. The number of chickens is twice the number of cows, and the number of spiders is twice the number of chickens. How many chickens are there?

“I set up the variable for cows as x, then chickens = 2x, spiders = 4x.”

View Post

SIG (Susquehanna)·Data Scientist·Technical Phone Screen

May 2026
Interviewed for a Data Scientist role at SIG and got hit with a classic knights and knaves logic puzzle. Pretty representative of the kind of lateral thinking they seem to value there.
  • On an island of knights (always truthful) and knaves (always lying), person B says to person A: 'You always lie.' Which role assignments are logically consistent, and why?

“I stared at this for a second because it looks simple but you have to be careful about each case.”

View Post

SIG (Susquehanna)·Data Scientist·Technical Phone Screen

May 2026
SIG (Susquehanna) threw a logic puzzle at me for a Data Scientist screen and it was the kind of thing that looks clean on paper until you're actually in it. One question, pure reasoning, no coding.
  • Three people (Asta, Bronya, and Clara) each have a different number of cats. Each person tells the truth when speaking to someone with fewer cats than them, and lies when speaking to someone with more cats. Given four statements they make to each other, and knowing not all statements are true, figure out how many cats Asta has.

“This one took me way longer than I'd like to admit.”

View Post

SIG (Susquehanna)·Data Scientist·Technical Phone Screen

May 2026
SIG had me work through a pretty dense stats and experimentation problem set built around a delivery platform scenario. Five sub-questions covering sample size math, non-inferiority testing, multiple comparison corrections, cluster randomization, and sequential monitoring. No metadata on outcome or timing was shared.
  • Given a baseline mean delivery time of 42 minutes with SD of 15 minutes and a target relative improvement of negative 5%, compute the per-arm sample size for a two-sample t-test using alpha 0.05 two-sided and 80% power.
  • For a non-inferiority test on cancellation rates where the baseline is 6% and the acceptable margin is 0.5 percentage points, compute the per-arm sample size using one-sided alpha of 0.05 and 80% power.
  • You have one primary metric, two guardrail metrics, and one secondary metric. Propose and justify an error control strategy across these metric families and explain how you'd report adjusted confidence intervals.
  • Randomization is at the zone-day level with 300 orders per cluster and an ICC of 0.03. Compute the design effect and the adjusted per-arm sample size, then figure out how many zone-days per arm are needed.
  • With four equally spaced interim looks using O'Brien-Fleming alpha spending, explain how early stopping boundaries compare to the final look boundary, and describe how this affects minimum detectable effect and trial runtime in practice.

“The mechanics here are standard: delta is 0.05 * 42 = 2.1 minutes, then plug into n = 2 * ((z_alpha/2 + z_beta) / delta * sigma)^2.”

View Post

SIG (Susquehanna)·Data Scientist·Technical Phone Screen

May 2026
SIG (Susquehanna) threw a classic river current brain teaser at me for a Data Scientist screen. One question, math-heavy, and they just watched how you worked through it. Not a vibe interview.
  • Two friends paddle upstream for 4 hours, then downstream for 5 hours. The next day they canoe back to their starting point, which is 23 miles upstream, and arrive at 4pm. The river current is 2 mph and their paddling speed relative to the water stays the same both days. What time did they leave on the return trip?

“I set up variables pretty quickly: let p be paddling speed relative to water, so upstream speed is p-2 and downstream speed is p+2.”

View Post

SIG (Susquehanna)·Data Scientist·Technical Phone Screen

May 2026
SIG (Susquehanna) threw a classic knights-and-knaves logic puzzle at me for a Data Scientist screen. Pretty pure reasoning, no coding, just work through the cases and defend your answer.
  • On an island where Knights always tell the truth and Knaves always lie, person B says to person A: 'You always lie.' Which of the four possible role assignments (Knight/Knight, Knight/Knave, Knave/Knight, Knave/Knave) are logically consistent?

“I actually liked this one.”

View Post