RSS Feed

The Joel on Software Discussion Group (CLOSED)

Is anyone here a Developer in the Quantitative Finance field, or anything else in the NY Financial Markets. (Trading Software, Risk Management etc.)

How high do the salaries go in NY and what is required to break in? Series 7, CFA, advanced degrees?

I ask because I hear very good numbers being easily possible in NY. If thats true are the hours insane?

I've actually just started looking into this, I also hear a lot of money is being thrown around in wall street right now.  I have some background in finance verticals, but not much of Risk Analysis, etc.  What specific domain knowledge is required anyways?
I'm very curious too to see what it takes to break into this market.  Is it enough to have been doing Java/J2EE for 9 years, and throw some finance buzz words on there to get an interview?

Would love to see some discussion on this and post some links about this.

Do you have a phD?

"Quants" (those who do quantitative finance) are usually different from developers in the IT department.  Quants usually have PhDs in fields which require lots of math: physics, math, sometimes computer science.  Lately several 'Quantitative Finance' or 'Computational Finance' Master level degree programs have started (Columbia, CMU, NYU, etc.)

From what I have observed, Quants usually design mathematical models and send them over to the developers to either implement them in code or to integrate them with existing systems.  This is a fuzzy line, sometimes quants do write code.

I am more familiar with the developers in the IT department.  My guess is that you will get the highest salary if you can deal with real-time data.  In other words, you should be able to take data coming off the network (tcp/ip/jms/tibco/whatever), process it, cache it, query it, etc.  You should be familiar with programming with threads. You should also have a fairly good understanding of databases and be comfortable with the Unix command line tools.

If you are contemplating a job offer for something called algorithmic trading...take it.

There are just a huge number of J2EE 'experts' out there.  My boss is actively hiring so we get to see lots of resumes.  It seems that J2EE is so huge that people who claim expertise in it don't have time to study core Java.  If a job specifically advertises for J2EE (and there are lots of those), you are in luck...  These projects are usually not very close to the tradnig desk, frankly the nature of these projects doesn't appeal to me.

To be perfectly honest, I don't know what hiring managers think of those who are not familiar with the financial industry already (assuming you are not a fresh graduate). 
If you don't know much about the financial markets, make sure you don't pretend to do so.  Wallstreet programmers usually know quite a lot of detail about the business ... they'll figure out your bluff.  Just tell them that you are interested in getting into the industry ... no big deal.

You should expect to work more than 40 hour weeks.  Sometimes a lot more, sometimes less so.

You bonuds will very likely be higher than what people get in other industries.  But don't get too excited by articles claiming $100K bonuses or 100% of your salary ... those articles are not talking about developers :)

Starting at a small trading firm is the best way of learning the business quickly.  Large firms provide the best benefits (sometimes that includes on-site gyms, baby sitting services, etc.)

If you get to work with traders, do so.  They will be loud, demanding and won't understand a word you say about programmning issues, but salaries and bonuses are highest the closer you are to the trading desk.

In order to actually start the process of job hunting, contact recruiters.  I've never heard of people sending their resumes directly to firms ... try dice, hotjobs and specifically mention that you want to get into this industry.

Good luck!

"If you get to work with traders, do so.  They will be loud, demanding and won't understand a word you say about programmning issues, but salaries and bonuses are highest the closer you are to the trading desk."

Do you work on a desk falcon?

No!! Don't get tempted by Wall St money!  It's not worth throwing your life away!

Well, some of the IT groups are nice and there are some smart people.  But most technology groups in investment banks are staffed by not-very-bright people trying to do software and claiming to be financial experts.

If you're a math expert and know programming, there might be some good opportunities, but hours will be brutal. No joke!

I've worked for both financial & software companies, and working for a software company is much better.

Better is very subjective. I want to get a feel of numbers plus lifestyle and see what the tradeoffs are.

I dont have a Phd. I would be willing to do a MS (if thats what it takes) and work with traders for now and afterwords if it seems like the money is good.

I was initially hoping regular exams like Series 7 might be enough.

Software companies tend to cap for SR Dev/Architects at 130-150K perm.

Now thats not bad money, but if I could get more, say 200K+ for some hard work now then it seems worth it.

I was mostly wondering if anyone had any real experience in these types of jobs. Maybe its not worth it.

>> Is it enough to have been doing Java/J2EE for 9 years, and throw some finance buzz words on there to get an interview?


Try that for a business-side 'pure' quant job and you'd be laughed out of the room... if you even got that far. More than likely you'd be laughed at repeatedly by any headhunter/agent you approached, if they even bothered to reply to you at all.

You might have a chance as a 'quant-developer' using the above approach, but you'd have to find a company with some very, very gullible interviewers.

The core quant-dev skillset is c++ and lots of maths. For a pure quant, a PhD in a numerate subject (maths/physics etc) is normally required along with the c++ skillset.

I work as a quant-dev btw. Pay is better than normal dev work, lower than pure quants/traders.

For people with no financial background, I think the best route into finance is to be a realtime/threads guru and shoot for a position writing trading applications. Many of the people hiring for these positions don't seem to care whether you can spell NYSE. They want superstar programmers and (say) they will teach you whatever they think you need to know about trading etc.

In most quant shops there are modelers and developers, with overlap as somebody already said. You will not be considered for a modeler position without a PhD in physics, math, finance, etc. Well, the one caveat might be if you have been trading a strategy successfully somewhere else, then the almighty dollar trumps all other considerations :)

To be a developer, it helps to be lucky. I got in because I had the right skillset at the right place and time, and I interviewed well enough that I was given a shot. Getting your CFA might help show that you're serious. However, the CFA is one of those Catch-22 certs: it takes 2-3 years, and one of the requirements is to be working in the industry. I've never met anyone who did a Computational Finance Masters. The one at Columbia looks like a really great program; I would probably do it if I had more time and money.

So in short, the barriers to entry are fairly high. You can see why the pay is above average, and why people who are "in" are reluctant to leave despite the workload, hours.

BTW, we really work bankers' hours and just tell all those horror stories to scare away competition :)

>> Is it enough to have been doing Java/J2EE for 9 years, and throw some finance buzz words on there to get an interview? <<

Here's an list of derivatives-related financial buzzwords. You should learn to discuss at least the basics of these before going for an interview.

* Risk neutral valuation
* Arbitrage-free pricing
* Futures contract pricing
* Put-call parity
* Moneyness
* Option time value
* Black-Scholes
* Black model
* Binomial options model
* Monte Carlo option model
* The Greeks in general
* Volatility and implied volatility
* Volatility smile
* Short rate model
* Hull-White model
* LIBOR Market Model
* Heath-Jarrow-Morton framework

Here's a list of maths-related financial buzzwords. You should learn to discuss at least the basics of these before going for an interview.

* Calculus
* Differential equations
* Linear algebra
* Numerical analysis
* Probability distributions
* Expected value
* Value at risk
* Risk-neutral measure
* Stochastic calculus
* Itô's lemma
* Fourier transform
* Girsanov's theorem
* Radon-Nikodym derivative
* Monte Carlo
* Partial differential equations
* Martingale representation theorem
* Feynman Kac Formula
* Dynkin formula
* Stochastic differential equations
* Stochastic volatility
* Numerical methods

Let's get real here. If you are really great with advanced mathematics and know about finance, you're set. You can take the $200,000 job and hopefully bank some serious cash before you burn out.

If you don't know about this stuff, and you're not a college sophomore, forget it.

And if anyone thinks Mark is just throwing out a bunch of scary looking terms to intimidate you, you are wrong.  I'm at a finance job at a small software company, and have brushed up against half of those, including Ito's lemma (which is at the core of stochastic calculus), martingales, volatility surfaces (with smile), and arbitrage-free monte-carlo rate simulation.  And I just work for a company that works for mortgage banks, about as far from the trading floor as you can get.

If you are serious about getting into quant, don't worry, it's not quantum mechanics.  It is a lot of big names to concepts that aren't horribly difficult.  It's a very interesting sort of applied probability in an environment with a huge amount of randomness.  There are a lot of pieces of knowledge that depend on each other, and the books are terribly impractical in their approaches, for the most part (they are written by PHDs, after all), so learning it does take some time.

And when it comes time to program with it, you'd be amazed how hard it is to get all the details of a simple frickin' interest rate swap modeled right.

BTW, is there any hope for a programmer who has been exposed to a lot of these concepts to squeeze their way into a junior quant job?

QED, JC, and Mark -

You guys seem to know quite a bit about the business, so I'd like to ask your advice.  I've always been intrigued by this area of programming, because it sounds interesting and, well, lucrative.  I don't have the credentials you've mentioned, though...

I did a math undergrad and an M.S. in Industrial Engineering/OR.  No Ph.D, though.  I've taken many but not all of the math courses Mark listed, but I'm definitely unfamiliar with the Finance terms.  The work I do is more on the optimization side (linear and non-linear optimization) rather than the stochastic side, and it's in manufacturing, not finance.  I took a boatload of the stochastic courses in grad school, but that hasn't been a big part of my job.  Most of my professional programming experience has been Java (Ruby/Python and so forth for my own side projects).  I haven't written a line of C++ since... oh man, I think it was 2000 or so).  I'd be willing to switch back, but I'm well aware that it would take a while to get up to speed with C++.

What do you think, would I have an angle in here?  Or is another degree (ie.,  PhD or maybe another MS in Finance), or more C++ pretty much a deal breaker?  Also, is NY by and large the only place you can do this, or is there a way to get in on the west coast?

>> What do you think, would I have an angle in here?  Or is another degree (ie.,  PhD or maybe another MS in Finance), or more C++ pretty much a deal breaker?  Also, is NY by and large the only place you can do this, or is there a way to get in on the west coast? <<

For a role as a quant developer (as opposed to quant modelling), you should be fine. You can either try to go direct to the trading floor (you might need C++) or work your way via the risk management department.

I'm in the UK, so I can't comment on locations in the US.

Thanks Mark -

In your experience, is a Ph.D an absolute must for the modeling roles?  Or are there MS's in financial engineering doing this?  I've looked into Berkeley and Columbia's programs, and they seem pretty good (and can be completed in a year or so).  Ph.D's in math are definitely smart, and I can see why people would want to hire them for this role, but I can't imagine doing a PhD in physics or math specifically to become a quant.

>> In your experience, is a Ph.D an absolute must for the modeling roles? <<

I don't have any experience with quant modelling, so I can't say for sure. My gut feeling is that it's not essential, and that you should do fine with a lower degree.

In reality, the purpose of a degree in this field is to get you to the interview stage - a way of culling resumes. Once you've reached the interview stage, the degree (or lack of degree) doesn't matter.

re. PhD or not for a 'quant' job -

I think this largely depends on who is hiring. I've seen areas in the UK where a 'PhD mafia' definitely exists, and who laugh at mere non-PhD mortals like us. Teams like this were entirely populated by Maths/Physics PhDs. I've also seen people get quant jobs with MScs in Financial Maths, so it is possible.

One area where you may have luck is outside of London/NY - ie Tokyo/HK etc... there is an abundance of good candidates for business roles in London/NY, whereas I've seen quant teams in Asia staffed almost entirely by non-PhDs, or PhDs in 'unusual' subjects eg Chem Eng.

It's also worth drawing the distinction between the various flavours of quants... there are more academic quants eg research quants who generally have better 'formal' qualifications, there are often 'desk quants' who act as part of a desk team, and various flavors of 'quant-devs', 'model-validation quants' etc who tend to get their hands dirtier with implementation details

The 'IT quant-dev' to quant route is possible, but people I've seen succeed here have done extra studying while doing quant-dev work, eg part-time Fin Math MScs as I mentioned above.

C++ is definitely the lingua franca of the quant world (not saying I agree with this btw, its just an observation). Expect to have to deal with a lot of proprietary, legacy code, and potentially lots of excel add-in style development if you are trader-facing.

I'm about to finish an MS in physics from an Ivy school, have been working in engineering for a fortune 500 for ~4 years (about 28 y.o.).
Good at math and programming; I took a stochastic calc course last fall and did well.

My ideal position would involve a mix of math and programming...I don't expect to be developing new models (since I don't have a research background/PhD) but I know that plain, vanilla programming in an IT department wouldn't hold my interest.

How many jobs are there in that "middle ground" where you use both? Any insight is appreciated.