Career in quantitative finance vs data science/technology

Joined
10/16/12
Messages
16
Points
13
I recently graduated from a top (Ivy League caliber) university in the US with a major in industrial engineering and operations research and a minor (almost enough courses for a major) in computer science. I have worked in finance for internships and full-time (including quantitative research at a major asset manager and fixed income research at a bulge bracket bank). I was laid off from a recent position and am considering an MFE as well as a career switch to data science/technology.

I am very interested in graduate programs in data science/statistics/machine learning - the tech industry is booming now and the financial industry continues to struggle. Would it be a smart career choice to move to data science? I believe I have a strong enough background in optimization/stats/computer science to excel in this type of work. I also think tech provides a more direct value-add to society, and finance continues to suffer blows to its reputation and poor standing in the public eye.

What are your long term views on tech vs finance?
 
You've already got the smarts to learn to invest your own money for you. And you've already done some finance and are already thinking about leaving? What does that tell you.

Do what interests you.
 
I can utterly guarantee you that earning more, does not equate to happiness if you do not enjoy the work. You'll still hate your work, you'll just be bribed better to keep doing it. And especially, when that comes with a work-life sacrifice, you very soon start to look forward to long weekends more than payday... even if you spend half of them just sleeping in catching up ;)
 
Sure, but one can find interesting jobs in multiple fields, it is not necessary to confine yourself to just one area, in my opinion.
 
I work in the 'data science' group of a marketing research firm now. I did a MS in Math Finance and graduated a year ago.

Working at my job for the last year has been fun. I rarely work overtime and work from home often. My salary is undoubtedly less than if I had landed an offer from a top place in finance. On the other hand, I enjoy my work, have free time and do not have a high COL, as I do not have to live in a metropolitan area.

Most of what I do is IT related. Largely, it's programming and configuring software and databases. I do a small amount of modeling as well. There are people around me who are more modeling focused - many types of regression and machine learning models are used. There are also people I work with who are more connected to data collection, i.e. sampling issues in the surveys we run.

A fundamental knowledge of databases and statistics is all that is really needed to do this work, however many people have higher degrees.

Hope that helps.
 
Thanks Andy.

My job is located in Horsham, PA.

I was contacted by a recruiter to interview for this job, so I'm not sure if that helps. I do believe the personal website I had set up as a student (it's gone now) was helpful for them to find me.
 
Jose, thanks for sharing your perspective. Can anyone comment on the availability and stability of jobs in finance vs data scientists? Several studies have projected a shortage of people with the skills of analyze massive amounts of data in the upcoming years.
 
Thanks Andy.

My job is located in Horsham, PA.

I was contacted by a recruiter to interview for this job, so I'm not sure if that helps. I do believe the personal website I had set up as a student (it's gone now) was helpful for them to find me.
I have a similar website in data analysis as I am looking to work more permanently in this sector. For the moment I'm selling the traffic, which is at least in the short term paying ok. But a little intense. I also have a call centre job on top of that.

Quite a few recruiters and employers are looking for this as it cuts a lot of crap out for them. Seeing PhD beside your name, or even cold calling employers doesn't say the same as "here's a dataset, here's a few analyses and here's why they're useful". In fact even top grads in finance or stats will make dumb errors thanks to no practical experience. The classic in quant finance is taking volatility out the BS equation - many will not spot it's a non-stochastic differential equation, something that a more practical candidate will spot.

To the OP - there's a few issues here. The reason I'm in this situation is because I didn't move when I had the chance. I used to be a quant, but my team got shut down right after I started. It took me 1 year to realise it, but I realised the stupid role they moved me into was A) not me and B) fell very far short in terms of mathematics (occasional excel is not even remotely close to using my skillset and left me uncompetitive in any mathematical job market). But by then I was locked into a completely different market which I wasn't suited to. I wish I understood it better, but what had happened is that too many of the quant skills I picked up from uni or being a quant were lost by then and employers know a grad or someone more recently employed in the sector would have them.

I stayed in PF for 6 years, but by the end I was dragging myself into work on a daily basis for a job that had very little to do with maths and which required numerous skills I shouldn't have been expected to have given my background and where sacking or being paid off was inevitable - I got paid off after that. For a few years the job had been ok as I got somewhere and moved to a far better firm, but ok became underperforming as frankly I was in the wrong job all along. And take a hint from Andy's question - he's the owner of this site and has to query how people made career changes? The reason even he doesn't know is because maths has become by some way one of the toughest areas to do career changes within - I was roadblocked every which way whenever I tried to correct this. So you will have to be careful. It's not impossible, but most people in other careers don't realise the implications and assume it's a career for life and that all this career change stuff is people being prima donnas - job loss/sacking/unemployability never comes into their minds given that you'll have plenty of other areas where people can either spend 40 years in the wrong job and not lose it, or swan into a new one with very little hassle. Actuary is the only mathematical career where you can take the piss that I know of, but even then it'll only take one major shift to change that. And there's indications even that's changed. Do you honestly want to bank on your position being protected by necessity for 40 years?

The thing to look at is day to day work. This is where things vary hugely - the previous poster describes his work as IT work, but that's quite ambiguous as I think he means programming and scripting. Most IT work has little programming and is more involved in servers/support for issues people have with their systems. I'm not gonna say which will be better for you - but be aware of the demands of this job and also don't try use it as a stepping stone to data science (in fact you might rule out data science by taking such a job - there no shortcuts or "stepping stones" and you won't have the energy to correct that mistake). If you like the adrenaline rush and can do this, maybe try it. But if you really prefer programming, avoid. There's a VAST difference between doing a job like the previous poster's (where it sounds like it's not exactly the dream you might expect but where his skillset is very much used and fun) and doing something like what I ended up in or IT support. Hell being quant wasn't all I expected it to be, but it was fun and the satisfaction and motivation I got (vs the incessant stress of other roles I did) tied in with it being the only job I ever had where I had virtually no issues performance wise. This proved hands down that every douchebag that came out with this "ah sure all jobs are the same" crap were patronising, condescending idiots and badly, badly wrong.

Also to round off the misery if you do wind up in the wrong job look forward to everybody in your life being self appointed experts in whatever it is you do (while in their tiny little world you know nothing), usually being wrong and pissing all over you, until you put it right. And putting it right can take time. It went wrong for me in 2006 and I'm still paying through the nose for it.
 
Last edited:
Jose, thanks for sharing your perspective. Can anyone comment on the availability and stability of jobs in finance vs data scientists? Several studies have projected a shortage of people with the skills of analyze massive amounts of data in the upcoming years.
It's definitely real. I can only comment anecdotally - I go to a lot of conferences where guys from Cloudera/Apache talk and things like using Hadoop for companies where data is exploding in size is something that was speculated about in 2013 but being used now by different businesses.
I'm not so sure about finance, but I think a lot of the new jobs in finance are in MBA stuff like M&A and loans - you don't need quants in those areas. In fact, until a fancy new product is developed within finance, which is where quants would be needed, becoming a quant may remain very tough indeed. And also there's too many MFEs, even allowing for the fact that some might be light on their skills e.g. how many of those programs are in touch with the market?

Recruiters probably have more ability to get people like Jose started within Data Science than finance as I've noticed most contact is answered proactively (indeed some have given some practical career change tips) while virtually any such request with finance recruiters is ignored. Until 2008 such requests to financial recruiters resulted in them practically dragging you from your desk, as then they could place you within weeks, now even the most experienced and talented are ignored.

Given that my experience is all within finance, it should be other way around as recruiting me within finance should be infinitely easier than something I've never worked in. This speaks volumes about which is in a better state.
 
Last edited:
I am very interested in graduate programs in data science/statistics/machine learning - the tech industry is booming now and the financial industry continues to struggle.

Are these topics future-proof?

In my experience technology changes dramatically every ~ 7 years.
 
I am very interested in graduate programs in data science/statistics/machine learning - the tech industry is booming now and the financial industry continues to struggle.

Are these topics future-proof?

In my experience technology changes dramatically every ~ 7 years.

Good to see someone has a brain here - I often state this and get some stupid response like "but technology is booming now". It seems to be a wildly revolutionary idea to some that markets move over time.

The trick is probably to not get caught up in the technology itself but use it as a platform to develop skills that will be valued in future.

E.g. with big data technologies like Hadoop focussing on the maths processes and the ins and outs of choosing when to use R or C++, rather than the ins and outs of firing up clusters, makes sense. Hadoop is a prime example where I can see something else coming along that is far simpler, whereby Hadoop specifics are suddenly useless but whatever replaces will probably be designed for data scientists in mind.

In fact in the way I'm seeing Hadoop developing this is already happening - the technology has been changed from a complicated Java based software to something where traditional data analysis skills are simply ported in. This means being an expert in things like managing clusters might have been lucrative in 2011, but now that most Hadoop programs companies use do it for them, that skill's potentially less in demand. But being able to use data analysis skills with Hadoop is not only unaffected but more sought after as Hadoop is now more geared towards having languages like R ported in. I think even if Big Data itself fell through this still holds true - surely an expertise in data analysis is more portable (albeit there would be hassle career changing) than knowing how to operate the machinery of Big Data.
 
Last edited:
surely an expertise in data analysis is more portable (albeit there would be hassle career changing) than knowing how to operate the machinery of Big Data.

That sums it up pretty well IMO.

Domain savvy >> product/technology knowledge.
 
surely an expertise in data analysis is more portable (albeit there would be hassle career changing) than knowing how to operate the machinery of Big Data.

That sums it up pretty well IMO.

Domain savvy >> product/technology knowledge.
I've volunteered in schools and wish the stupid promotional material they have got this across. Maybe some quants or bankers that now teach maths in private schools get this across, but usually some HResque idiot careers advisor gets listened to as they "sound" more business savvy. Ironically what you've described is what works best from a business perspective, not their often contradictory crap.
 
Last edited:
In fact in the way I'm seeing Hadoop developing this is already happening - the technology has been changed from a complicated Java based software to something where traditional data analysis skills are simply ported in. This means being an expert in things like managing clusters might have been lucrative in 2011, but now that most Hadoop programs companies use do it for them

Eventually companies like IBM will take over this space, Hadoop will disappear and the market will be sown up and become a commodity. Until the next new wave; jump in before the rest and get out when it starts to decline.
 
In fact in the way I'm seeing Hadoop developing this is already happening - the technology has been changed from a complicated Java based software to something where traditional data analysis skills are simply ported in. This means being an expert in things like managing clusters might have been lucrative in 2011, but now that most Hadoop programs companies use do it for them

Eventually companies like IBM will take over this space, Hadoop will disappear and the market will be sown up and become a commodity. Until the next new wave; jump in before the rest and get out when it starts to decline.
And hope you get the right experience along the way which can make the "getting out" bit easy.
 
An update from nearly 9 years later:

I appreciate everyone who commented on the thread and gave their thoughts. At the time I wrote the original post I was young, anxious, a bit lost, and much more career-obsessed than I am now. I attended one of the MFE programs ranked in the top 5 on this site, which I enjoyed. After graduation I worked at buy side firms focused mostly on quantitative alpha research, most recently for a well-known statistical arbitrage hedge fund. At the fund I was the only MFE and one of the few non-PhD researchers.

I currently work in data science/machine learning at a tech unicorn, so have tried both the tech data science and finance jobs. In many ways the jobs are more similar than I thought. In both my quant group and DS group, I collect data, build models using statistics and machine learning, and write production software. I have found the tech world to be more friendly and relaxed while also paying well, which has made it a good fit for me. While it has been several years, I still get unsolicited calls and messages from finance headhunters usually multiple times a week. While I plan to stay at my job for now, I would consider taking another finance job, but only at a place with a very strong research and tech culture and direct impact on PnL. For the same compensation, I would choose tech. Finance would need to provide a substantial compensation boost to make up for the additional stress, employment risk, and volatility.

Hopefully my story is helpful to others considering similar decisions. Both quant finance and data science/ML/AI have progressed since I was in school. Most of the top quant funds do not hire from MFE programs, and while I enjoyed my program and my classmates and the investment more than paid off, I am not sure I would do it again. Many MFE programs traditionally focus on stochastic calculus, PDE, and derivatives pricing, skills for which the demand has been stagnant at best for the last ~15 years. Core math, statistics, and computer science knowledge is durable over time and portable to many different jobs and in my opinion provide great optionality. In both quant and DS roles I have found joy and excitement in the discovery of market anomalies or development of a new ML models. I also enjoy the craftsmanship of building quality software. In all cases curiosity and persistence has made a great difference.
 
Back
Top Bottom