Mathematical Biology to Quant Finance

[almost_uncertainly]

Hello. Thanks in advance for any advice.

I was wondering if anyone had any advice on making a transition from mathematical biology to quant finance.

I have been involved in making stochastic models to investigate biological systems for PhD and postdoctoral research. The tools I've been using are relatively diverse, ranging from discrete time models, to stochastic calculus. I'm comfortable dealing with SDEs both numerically (primarily using Python, but also R and Matlab) and analytically.

I have recently decided to leave academia, and the logical continuation to me seems to be a quant role, partly because there are overlaps with my existing expertise. I had a few questions which I was wondering about:

1) Is it worth taking a Master's in Quantitative Finance? I am happy to do so, but I wonder if going back to do a Master's is overkill given that I already have a PhD.

2) Is age an issue towards securing a position?

3) So far I have had little success in direct applications towards firms and also no success with recruiters. Is it possible that I am phrasing my CV too academically? Should I be opting for a more commercially pitched CV?

Thanks for any pointers.

 

[cab]

I can't completely answer your questions, but I came from Math Bio and am in a MFE program now, so can add a little bit of prospective. I am still a student so take this with a grain of salt.

My undergrad degree was in math and physics, and I did a master's level project on stochastic versions of the Hodgkin-Huxley model which involved some stochastic calculus and a lot of numerical work. I, like you (presumably), looked up mathematical finance on the web and saw a bunch of stochastic calculus. Since I had an fairly exceptional exposure to this area I thought that my pre-existing skills would be relevant to the quant finance world.

However, now that I have gotten more perspective on the industry, through talking to people, going to presentations, interviews, and networking, I don't think stochastic calculus is as pervasive, or useful as it might appear. From what I gather, stochastic calculus is used very rarely on the buy side (I assume you know the buy-side, sell-side distinction) which is dominated by machine learning and statistics. Stochastic calculus is used on the sell-side, but there are three things to note:

a.) Most sell side jobs use it sparingly
b.) The jobs that do use it a lot are not so numerous, and are very competitive
c.) There are only 10-15 recognizable companies to apply to on the sell side
d.) Exotic derivative pricing (the main application area of real StoCal) has tapered off significantly in investment banks since regulations imposed by the Dodd-Frank act

To be more specific, I remember giving my, "I have experience with the development and analysis of SDE models" pitch to two investment bank quants and received a, "...we don't do that as much anymore" kind of response in both cases.

My main point here is that (from my experience) stochastic calculus is not the silver bullet which gets you into these banks/shops.

In my opinion, to be successful getting a job as a quant you need to have a wide range of skills to market, so that you are employable in the greatest number of jobs. The SDE heavy jobs represent like 3% of the quant job market (this is a B.S. guess but I have some confidence in it) , so if you apply to anything outside of the 3%, talking a lot about SDEs signals that you are trying to be in the 3% and are applying to the wrong job.

You have certainly done a great deal of stochastic data analysis, scientific computing, and mathematical modeling (Statistics is also very important if you have some working knowledge there). Consider marketing yourself using these more general skills, instead of the focused ones that you presented in your original post, and I think you will grab the attention of more people.

Also, I can't say whether an MFE is right for you, but I am pretty sure that if you get into a good one, that you will get a good job (since you already have done substantial research). My impression is that most of the initial hiring is done through target schools, and that it is hard to get in outside of this circle.

Oh, and I don't think it is unreasonable to get another degree if you are transferring to a completely new field. There are a lot of new things to learn, and you get to use the network of the school. The question is whether it is worth the hefty tuition price tag.

best of luck

 

[Daniel Duffy]

Hello. Thanks in advance for any advice.

I was wondering if anyone had any advice on making a transition from mathematical biology to quant finance.

I have been involved in making stochastic models to investigate biological systems for PhD and postdoctoral research. The tools I've been using are relatively diverse, ranging from discrete time models, to stochastic calculus. I'm comfortable dealing with SDEs both numerically (primarily using Python, but also R and Matlab) and analytically.

I have recently decided to leave academia, and the logical continuation to me seems to be a quant role, partly because there are overlaps with my existing expertise. I had a few questions which I was wondering about:

1) Is it worth taking a Master's in Quantitative Finance? I am happy to do so, but I wonder if going back to do a Master's is overkill given that I already have a PhD.

2) Is age an issue towards securing a position?

3) So far I have had little success in direct applications towards firms and also no success with recruiters. Is it possible that I am phrasing my CV too academically? Should I be opting for a more commercially pitched CV?

Thanks for any pointers.

Knowing just a bit/a lot of stochastics is not a reason why you can make a transitions from biology to finance.

You need to research the topic in more depth.

 

[mhy]

1) A MFE/MQF may help you get in front of recruiters. It will also help with your knowledge. Examples of questions you will be able to answer (or at least know how to research to find the answer) after an MFE: How do you interpret the treasury curve? How do you price options and derivatives? What are the commonly used interest rate models? What does 'trading vol' mean? What is a 'spread trade'? How do you interpret economic reports? How do you deal with noisy financial time series? How do you perform Monte-Carlo simulation on a stock price to estimate the theoretical option value?

2) No, but some firms do prefer young grads so they can 'mold' them.

 

[pingu]

From what I gather, stochastic calculus is used very rarely on the buy side (I assume you know the buy-side, sell-side distinction) which is dominated by machine learning and statistics.

It depends on what they buy-side is buying. If you are dealing with mortgages or structured products, it will come handy.

 

[Liam]

1) A MFE/MQF may help you get in front of recruiters. It will also help with your knowledge. Examples of questions you will be able to answer (or at least know how to research to find the answer) after an MFE: How do you interpret the treasury curve? How do you price options and derivatives? What are the commonly used interest rate models? What does 'trading vol' mean? What is a 'spread trade'? How do you interpret economic reports? How do you deal with noisy financial time series? How do you perform Monte-Carlo simulation on a stock price to estimate the theoretical option value?

2) No, but some firms do prefer young grads so they can 'mold' them.

Not sure how old this thread is but point 2 is important. I really can't think of a more difficult job or career to get into if a) you're too old or b) you've done something else. Don't get me wrong, it's not impossible as I know a couple of programming to quant career changers but that needs to be considered against the hundreds of programmers or statisticians with stochastic based degrees that I know that have tried and failed - and we're not talking about people that haven't studied it in depth, we're talking people that probably go through Hull and all those sde books while taking expert advice on what the market wants.

It's nothing to do with competition, although I'd say the number of MFE grads is huge. The poster I've quoted has it in one - it's about moulding and market risk is similar. In fact in some cases many firms complain they can't get enough people that match what their looking for - regulation makes it even more difficult and it's about standards they need in candidates rather than some dumb filtering process.

Up to a point it's a case of get the quant thing right in one - this is a stressful task as going for something like quant roles has a very all or nothing approach. Don't just think "ah sure there's loads of other areas of maths" - failing to become a quant and then trying to find an alternative is hard (especially if you've done MFE which makes the alternative look like "Plan B").

 

[yetanotherquant]

I was wondering if anyone had any advice on making a transition from mathematical biology to quant finance.

Well, I know a guy, who successfully changed: http://www.hedgework.de/schon-wenige-tage-extremer-verluste-können-ganze-portfolios-an-den-rand-des-ruins-führen.html

Knowing just a bit/a lot of stochastics is not a reason why you can make a transitions from biology to finance.
You need to research the topic in more depth.

Yes, that's true. The times, as everyone with (some) math knowledge was welcome in quantitative finance are längst vorbei (far away in past).
More and more special knowledge is currently expected from quants.
Have a look at wannabequant advice: http://www.markjoshi.com/downloads/advice.pdf
And - if you wanna work in trading or portfolio management, at my book: http://www.yetanotherquant.com

 

[Daniel Duffy]

Well, I know a guy, who successfully changed: http://www.hedgework.de/schon-wenige-tage-extremer-verluste-können-ganze-portfolios-an-den-rand-des-ruins-führen.html


Yes, that's true. The times, as everyone with (some) math knowledge was welcome in quantitative finance are längst vorbei (far away in past).
More and more special knowledge is currently expected from quants.
Have a look at wannabequant advice: http://www.markjoshi.com/downloads/advice.pdf
And - if you wanna work in trading or portfolio management, at my book: http://www.yetanotherquant.com

What's new? This is herd behaviour, everyone saying the same things.
The economy drives everything In a few years things will be booming and people tend to pay less attention to detail, and not just in quant (which is a specific niche area).

All technology moves in cycles; you just have to know where you are on the curve. Amen.

" längst vorbei (far away in past)."
And in der Zukunft!

 

[Daniel Duffy]

mathematical biology

What is that?
ODEs/SDEs?? or more advanced?

 

[yetanotherquant]

>What's new?
>The economy drives everything ...
>All technology moves in cycles;
But it does not just oscillate, it also has a trend component
And the globalization is its name!
It is hard to compete (both in term of costs and math knowledge) with brilliant guys from China, India or Russia.
But one can compete them in terms of specific financial knowledge, e.g. regulatory standards (and their practical application).

Thus

Knowing just a bit/a lot of stochastics is not a reason why you can make a transitions from biology to finance.
You need to research the topic in more depth.

 

[Daniel Duffy]

On the other hand, the real financial _innovative_ maths is coming from Italy and France.

I don't see much coming from the former Soviet Union these days. The Cold War is over and maths level has changed for the worse.

BTW, how do you measure mathematical intelligence? Olympiad brainteasers?

 

[yetanotherquant]

On the other hand, the real financial _innovative_ maths is coming from Italy and France.

I don't see much coming from the former Soviet Union these days. The Cold War is over and maths level has changed for the worse.

BTW, how do you measure mathematical intelligence? Olympiad brainteasers?

Would agree for Italy (Brigo, Mercurio, Ballabio).
But France (which totally infected with Bourbaki) and financial math (which is about modeling a real and quickly-changing world) - its an oxymoron!

Yes, Russian level has degraded but Chinese has grown (American University is a place where Russian Profs teach Chinese students maths).

As to measure of mathematical intelligence, I just apply Lebesgue measure (which is only translation-invariant measure and the math knowledge is translation invariant, that's why so many mathematicians moved from exUSSR to USA )).
Well, to be serious its a matter of thinking culture but let's put it so: Perelman managed (and Hamilton not) to prove Poincare Ansatz because Perelman possesses the erudition of the Soviet math school.

 

[Daniel Duffy]

Would agree for Italy (Brigo, Mercurio, Ballabio).
But France (which totally infected with Bourbaki) and financial math (which is about modeling a real and quickly-changing world) - its an oxymoron!

Yes, Russian level has degraded but Chinese has grown (American University is a place where Russian Profs teach Chinese students maths).

As to measure of mathematical intelligence, I just apply Lebesgue measure (which is only translation-invariant measure and the math knowledge is translation invariant, that's why so many mathematicians moved from exUSSR to USA )).
Well, to be serious its a matter of thinking culture but let's put it so: Perelman managed (and Hamilton not) to prove Poincare Ansatz because Perelman possesses the erudition of the Soviet math school.

RIP Soviet Science?

Lebesgue measure
This is a piece of almost useless maths (and yes, I studied it for 4 years as an undergrad).

Perelman managed (and Hamilton not)
Bravo.
Hamilton had better things to do.

People follow the money.

 

[yetanotherquant]

>RIP Soviet Science?
Not RIP but rip-off

>Lebesgue measure
>This is a piece of almost useless maths
Well, we have polar opinions on this.
And how about the book "Computational Complexity: A Quantitative Perspective" (yes, finally I found it ): www.amazon.com/dp/B00O4WDQWO/
It uses measure theory (and topology)

 

[Daniel Duffy]

>RIP Soviet Science?
Not RIP but rip-off

>Lebesgue measure
>This is a piece of almost useless maths
Well, we have polar opinions on this.
And how about the book "Computational Complexity: A Quantitative Perspective" (yes, finally I found it ): www.amazon.com/dp/B00O4WDQWO/
It uses measure theory (and topology)

It's not an opinion! It is a fact, just look around.
A nice price! This is an academic book.

Can you please give any _real_ applications of measure theory?

 

[yetanotherquant]

>Can you please give any _real_ applications of measure theory?
Of course. In LIBOR model you need to find (and compute!) this f*ucking change to T-Forward measure in order to use this model.

 

[Daniel Duffy]

>Can you please give any _real_ applications of measure theory?
Of course. In LIBOR model you need to find (and compute!) this f*ucking change to T-Forward measure in order to use this model.

That's an example, fair enough. Girsanov and Radon-Nikodym are also useful.
I suppose it depends on what one means by 'application'. For me, not.

Let's call it a draw.

 

[yetanotherquant]

>Let's call it a draw.
ok!