Jane Street · Primly Community

Jane Street interview rejection post-mortem, what I'd change if I did it again

mobile_mara · 6 replies

Got rejected after the Jane Street final round last month. Quant SWE track, 2026 grad. I've had a few weeks to sit with it and I want to write down what I think went wrong, partly for myself and partly because I couldn't find much honest post-mortem content when I was prepping.

What the loop looked like: OCaml coding screen, a second coding round focused on functional style, a probability/trading round, and a systems round. Four stages, multiple interviewers each.

Where I think I lost it:

OCaml. I knew it was coming and I still wasn't fluent enough. I can read it but writing under pressure in an unfamiliar functional style was painful. I was slow. They value fluency, not just ability to figure it out eventually. If you haven't shipped real OCaml code, start earlier than I did.

Probability round. I choked on a relatively standard expected value problem because I tried to set up a formal solution before I reasoned through the simple case. The interviewer gave me a hint and I recovered, but you can feel when the energy in the room shifts. The recovery probably didn't look great.

Didn't think out loud enough. Jane Street values how you reason, not just whether you get the answer. I got a couple of answers right but I was too quiet getting there. I think I lost points on process.

What I'd do differently: spend at least 6-8 weeks on OCaml specifically, not just reading docs but actually building something. Pair that with serious probability and market-making problems from the Jane Street puzzle archive. And practice narrating your thinking even when you're figuring something out.

I'm reapplying when the next cycle opens. No shame in that.

6 replies

bootcamp_bri

The OCaml point is something I see people underestimate constantly. It's not a language you can pick up over a weekend. Even if you're solid in Haskell you'll still be slower in OCaml at first. The idioms are different.

analyst_ana

How long was the probability round? And was it heavily market-making framed, or more pure probability puzzles?

jp_newgrad

45 minutes, 2 problems. The first was pure probability, the second was more market-making adjacent, like quoting prices given some information and explaining why. Not as structured as Leetcode, more free-form.

market_realist

Thinking out loud thing is so real at Jane Street. I did an internship there and the debrief feedback was basically all about reasoning process. Getting the right answer quietly is less impressive to them than getting a slightly wrong answer while showing great structured thinking.

intl_isla

Did they give you any feedback when they rejected you or was it just a no?

jp_newgrad

No real feedback, just the standard 'we won't be moving forward' email. Which is annoying but I get it, they probably do hundreds of these loops.