Ask HN: How to get a job at Jane St., Two Sigma, etc.? | Hacker News Reader

Ask HN: How to get a job at Jane St., Two Sigma, etc.?

I’m a full stack software developer with about 10 years of experience building things like APIs, mobile apps, various backend services, SDKs, etc.

I’ve long been interested in finance and am always mistaken for working in finance (guess I just look the part).

I’m wondering if anyone here has gone from the world of software for the stereotypical Bay Area startup or large tech company to world of NYC quants and hedge funds. I don’t have a PhD in mathematics or some other extreme traditional qualification, but I do have lots experience building stuff.

I recently moved to NYC and would love to join one of these companies, especially considering the earning potential versus other companies.

Has anyone here made that move?

View on HN

I worked for a few years in fintech (NYC) after I left SF (more than a decade working at FAANGs).

I'd say your experience as a full stack software developer is basically irrelevant. At the end of the day, for a typical non-leadership engineering role, the hiring decision is not going to be based on your previous work experience. They are looking for strong algorithmic fundamentals and the ability to think outside the box (I've interviewed for both in the past).

How were you able to get an interview at that company? Was it through networking or did your FAANG experience show some kind of algorithmic prowess?

I'm a university student who's found out a little too late that I might be interested in Quant. I have great marks and algorithmic foundations but my past experience has been in Full Stack/Web Dev, and I've struggled to get an interview from any of these firms (and I don't entirely blame them).

Maybe they're not looking for current students? That alone leads me to pass over a candidate unless I'm trying to fill spots for a summer intern program.

> I'd say your experience as a full stack software developer is basically irrelevant

This seems to basically be true for any well paying ( > $200K) job at any of the better paying tech companies in the US. It’s all about the passing the algorithm coding interviews.

> It’s all about the passing the algorithm coding interviews.

As someone who's been involved in hiring for these companies - it's more accurate to say that for all but the most exceptional candidates, a decent score on the algorithms interview is necessary but not sufficient. I've seen fairly similar performances in technical interviews result in offers for mid-level (~$350k), senior (~$500k), or staff (~$700k) based on experience and the behavioral interview.

Of course, every leveling guidelines for tech companies level above mid based on system design, and history of working problems involving “scope” and “impact”

I don’t believe that’s true.

There a number of large, well funded “startups” that don’t do the traditional algorithm based interviews but still pay at or above that range.

So what are these startups and do they pay in the range of these companies.

https://www.levels.fyi/2020/

Are you talking for a quant role, or a SWE role in those companies?

I can’t imagine how 10 years of experience as an SWE would be irrelevant for another SWE position, even if the industry is different.

As someone at a quant firm, it's true of both.

Years of experience really counts for next to nothing. Two people with X years of experience could have vastly different skillsets and it genuinely signals next to nothing about how qualified they would be for this job.

Both roles involve highly technical interviews and then a 3 month probationary period that is quite challenging. If you have a lot of experience then I'd expect you to do well in the technical interview, although unfortunately in practice this doesn't end up being the case.

Part of the question there is whether a candidate actually has ten years worth of experience, of one year of experience ten times.

I have. The hedge fund world isn't really as enigmatic as they would have you think. It's the same as anywhere else: work at a reputable company, get an interview, do well. I don't think it's much different than working at a FAANG. Lots of ex-Google employees at Two Sigma, for example.

Things are a little different if you want a job as a quant than a software developer, though.

Definitely want to keep a SWE role. I am effectively trying to find how I can best maximize my earning potential as a SWE.

Cool. Just a few thoughts:

The tech and engineering discipline are always going to be worst than FAANG or one of the darling startups. That's part of what they pay the big bucks for, to minimize attrition. The flipside is that you can have an outsized impact you probably can't at a FAANG. These are the qualities you want to emphasize to get hired from industry. Hedge funds hire a lot of green kids out of good colleges so you need to bring something to the table more than just being a smart person.

It's also probably not super helpful to you right now but the industry is pretty small. Once you've had success at one hedge fund it's pretty easy to get interviews at the others.

How much of a pay gap is there between the quants and swe at these firms?

It's highly dependent on performance since quants are tied directly to a PnL number. If that number is really, really good they're making millions. More average quants are still making a premium over senior SWEs, underperforming quants don't stay quants for long.

This all varies firm to firm though, especially depending on what kind of trading they do.

Just random thoughts:

Be pretty good at what you are doing (absolutely no need to be a genius though, at least for SWE). Classic Google interview stuff, although they do put emphasis on slightly different things.

Knowing maths helps, knowing low latency helps, but its not required - any quant shop needs a ton of normal software.

If you don't have friends in the industry, etc. going through a good recruiter might actually help.

> How to get a job at Jane St., Two Sigma, etc.?

Which "job"?

If you are going for a role as a quantitative trader, then your job will involve building mathematical models for predicting the price of assets. You would need to study up on probability theory, statistics, and risk management.

If you are going for a role as a software engineer, then your job will involve building internal tools and infrastructure for productionizing the models that quant traders build. You would need to study up on data structures, algorithms, and systems design.

The latter.

Thanks for clarifying. Depending on your job search strategy, you could apply as a generic software engineer or you could try to apply under a specialization. Specialists are valuable if the company needs you, but useless if they don't.

Here are some examples of specializations in quantitative trading:

Networking. For high-frequency trading firms, low network latency is extremely valuable. Some firms will need to build their own proprietary network protocols for their internal services because TCP is too slow.

Compilers. Again, for high-frequency trading firms, the latency from garbage collection, memory allocation, and system calls must be managed very carefully. Jane St is famous for being an OCaml shop, and OCaml is commonly used for building compilers and domain-specific languages (DSLs). When it comes to optimization, predictable latency (variance) can be just as valuable as low latency (99th percentile).

ML Engineering. While the quant traders build models, it is up to the software engineers to productionize and maintain the models. This can involve data engineering, feature engineering, model deployment, and monitoring the model for performance, predictive accuracy, and feature drift.

Info might be a bit dated, but, when it was a possible path for me, Jane Street was an OCaml house and did a fair bit of picking from Columbia grads who took the OCaml version of Programming Languages & Translators. Legend also says they would ask candidates to do math problems in their head while the interviewers crinkled newspaper and talked loudly. Fair to say they are looking for exceptionally skilled and sharp candidates.

Candidate: "I hope you don't mind me putting on some noise cancelling headphones while I work on this problem, please tap my shoulder if you have any questions."

interviewer (thinking: "Ah! Finally we have a proactive self-reliant candidate who knows his own strengths and weaknesses!"): "Don't bother, you're hired. Let's go to HR so we can work out the details."

Fair to say they have shitty hiring practices.

If you want to give someone an IQ test give them an IQ test, don't crumple fucking papers while you talk loudly and ask them math questions. Sounds like a power trip.

Would you like the job to be a painful surprise?

I've had nice cushy dev roles where I had all the time I wanted and other support-adjacent roles where I felt like I was working surrounded by screaming toddlers - expected to sit on an incident response bridge while diagnosing and fixing things.

I'm glad they told me what the role was like before I took it. The interview was friendly and casual, but involved a few people asking me to diagnose something and just cutting off "failed" paths and giving me the next bit of info, asking what that would imply and what we should test, to estimate how long it would take. If they'd de-stressed the interview the people they hired probably wouldn't be capable of or interested in the position. If they'd crumpled paper it would have helped by mimicking all the background noise from thirty unmuted phone calls.

I hear these roles often involve coding in a boardroom meeting with a few quants and a few engineers, going back and forth with the model and performance issues. If this was done by email it could take weeks and the opportunity could be gone. If you aren't able to keep hacking at your task while the other handful of people are enthusiastically discussing another then you aren't going to be able to help the team.

It’s a legal minefield to directly administer IQ tests in hiring because of all the legal precedent around them

It's easier than many people make it out to be. I've been given a couple of IQ tests during a hiring in process.

Not saying it's a good hiring practice, but in theory shouldn't interviews match the real job as closely as possible? Trading desks can be loud, frenzied environments to do work in and if that's an issue for people then it's better to know at the interview stage.

Yea, OCaml is still referenced on their job posts.

The math PhDs are for quants, sounds like you want to work as a dev. I think they still have very high standards for those, but it's probably quite similar overall to getting a FAANG job.

Depending on their strategy and stack they may have specific backgrounds that they look for for some roles, eg telecoms signal engineering, networking, maybe compilers or UNIX internals, but it's probably worth just giving it a shot with whatever your experience is. A finance background isn't required, they probably also won't care if you trade on your own. So I guess the advice is just practice some LeetCode.

Since you asked about Jane Street, they're famous for using OCAML, so that'd be a good thing to look into if you want to demonstrate a desire to work for them specifically.

I ran a data science group at a major discretionary NYC based hedge fund. We were not a traditional quant fund, but hired many software engineers.

Here is my take on what it takes: https://magis.substack.com/p/what-makes-alternative-data-sci...

Leetcode and possibly these problems: https://www.amazon.com/Heard-Street-Quantitative-Questions-I...

Reply on news.ycombinator.com