applying to two sigma new grad swe roles for 2026 and i'm kind of spiraling about how to prep. i'm a cs major, solid on data structures and algorithms, but i have basically zero quant or finance knowledge. should i be cramming options pricing or is that irrelevant for the swe track?
before people say "just do leetcode" -- i've been doing leetcode. i'm more worried about the parts i don't know i don't know.
some specific things i'm unclear on: their application says things like "strong mathematical background" -- do they actually test math in the new grad SWE interview or is that just covering their bases for quant roles? i've seen old posts saying they ask probability questions in the SWE loop. is that still true in 2025/2026? like expected value problems or conditional probability stuff? how is their OA structured? i did their Codility OA a while back and got through but wasn't sure if that's typical. is the behavioral component real or is it basically filler at the SWE level?
if anyone went through the new grad SWE loop recently i'd really appreciate any signal. i feel like i'm prepping in a vacuum.
4 replies
remote_swe_42
went through the new grad SWE loop two years ago so things may have shifted. but: yes, expect some probability basics. not deep quant stuff, but "expected value of this game" type questions, maybe a brain teaser. they want to see you can think probabilistically. you don't need to know what a delta is. also behavioral is real, not filler -- they care about intellectual curiosity and how you approach ambiguous problems.
jordan_pm
i'm in the same boat and have been freaking out about this. just saw a post from someone who went through it last fall who said they got a probability question in the first technical round. something about dice. not hard but unexpected if you only prepped algorithms.
consultant_cam
"strong mathematical background" in quant-shop job reqs typically means: can you handle quantitative reasoning, not just code? for new grad SWE it usually cashes out as: probability basics, maybe some stats intuition, clean algorithmic thinking. you don't need finance. but if you can't reason about expected value and conditional probability you'll get caught. spend 3-4 days on that, it's not deep.
pivot_pat
their OA format when i looked into it was 2 algorithm problems, 1 hour. one easier warm-up, one medium-hard. nothing crazy but time management matters. if you're Codility-comfortable you're probably fine on the OA.