I knew some quants like Dr. Wilmott have a background in fluid dynamics. Does any professor in financial engineering have a background in fluid dynamics?Especially does anyone know what?s the occupation of Daniel J. Duffy now, the author of the book ?Finite Difference Methods in Financial Engineering??I am just wondering what the career status of someone from computational fluid dynamics and finite difference should be.

Not sure what exactly you mean by fluid dynamics but check this page.

Daniel Duffy is Cuchulainn on this very forum...try asking him?

QuoteOriginally posted by: EdisonCruiseI knew some quants like Dr. Wilmott have a background in fluid dynamics. Does any professor in financial engineering have a background in fluid dynamics?Especially does anyone know what?s the occupation of Daniel J. Duffy now, the author of the book ?Finite Difference Methods in Financial Engineering??I am just wondering what the career status of someone from computational fluid dynamics and finite difference should be.You called, Sir? I am alive and (just about) kicking I have a background in Numerical Analysis(FEM,FDM). Which can be applied to many things, which I did/do.Self-employed, my own company since 1987 :)Disclaimer; I don't play golf.Thank you for your interest.hthedit: to be honest, people like PW and moi are few and far between in the sense that most do not study fluids or numerics. Just sayin' those subjects are not so popular anymore. Paul popularized PDEs in finance.

Great to see you here, Daniel. And thank you all for your replies.I have a background in PDE/FDM, fluid dynamics and found financial job positions related to these areas are so few nowadays. Though I am now in a financial engineer position in a developing country, related exotic products, barrier option, etc, attract very few clients. It is not that I cannot price and hedge the OTC product well, but that our clients cannot accept the product, because of the expensive price, the difficult concept, etc.. Anyway, our business is hard to go on. Previously I though OTC option is a good place to apply my PDE/FDM skills and generate remarkable financial return, but our country?s market may not need it. So I am considering if I should go on this path or go back to academic world or another industry. To continue or not to continue, that?s a question.

I'm sure big companies have research depts. where maths and engineering are needed. But I have no up-to-date info. Maybe things have changed. Some time ago I interviewed with both BP and Shell to develop FEM package for reservoir engineering from the ground up. Nowadays it is kind of open source. I do think there are opportunities for C++/C# (and Fortran) developers in this area. "Research" is not what it used to be.

Young people today don't even know the meaning of research!P

QuoteOriginally posted by: PaulYoung people today don't even know the meaning of research!PYoung people are very focused on a single outcome and it shows. The lateral thinking of say Polya is missing in general. Maybe they don't have the time.

QuoteOriginally posted by: EdisonCruiseGreat to see you here, Daniel. And thank you all for your replies.I have a background in PDE/FDM, fluid dynamics and found financial job positions related to these areas are so few nowadays. Though I am now in a financial engineer position in a developing country, related exotic products, barrier option, etc, attract very few clients. It is not that I cannot price and hedge the OTC product well, but that our clients cannot accept the product, because of the expensive price, the difficult concept, etc.. Anyway, our business is hard to go on. Previously I though OTC option is a good place to apply my PDE/FDM skills and generate remarkable financial return, but our country?s market may not need it. So I am considering if I should go on this path or go back to academic world or another industry. To continue or not to continue, that?s a question.I went through the same kind of reasoning you went through years ago, alas it is very naive. First of all a lot of pde engines have been written years ago, I mean years ago. Second pdes in derivatives aren't particularly interesting. I know there are a lot of people going crazy with all sorts of stuff here, but fundamentally it's just fairly plain and simple engineering. Moreover it is a very important but also very small aspect of a system.It's a little bit like splines in computer graphics and video games. It's kind of important, but it's burried among a thousand other important things that make a good game.

What I have noticed that nearly all PDEs here are linear. Second, Crank Nicolson and ADI are still used even though they are severely alt-modisch. For the record, we were fed CN in 2nd year undergrad and warned.In engineering 90% of models are nonlinear. Why not in finance? PW and Marco A. do deal with NL stuff. QuoteIt's a little bit like splines in computer graphics and video gamesThat's a bit harsh. It is more like hidden-surface removal and shading.. BTW try writing a 3d FEM engine for black-body radiation in a roundy tube:D

Hi Daniel,QuoteIn engineering 90% of models are nonlinear. Why not in finance? PW and Marco A. do deal with NL stuff.May I ask some details? What's the full name of PW and Marco A.? Thanks!

QuoteOriginally posted by: ShihHauTanHi Daniel,QuoteIn engineering 90% of models are nonlinear. Why not in finance? PW and Marco A. do deal with NL stuff.May I ask some details? What's the full name of PW and Marco A.? Thanks!Hi Shih-Hau,PW == Paul Wilmott, the Boss Marco Avellenada. See UVM type models, for example. HJB etc.

QuoteOriginally posted by: CuchulainnWhat I have noticed that nearly all PDEs here are linear. Second, Crank Nicolson and ADI are still used even though they are severely alt-modisch. For the record, we were fed CN in 2nd year undergrad and warned.In engineering 90% of models are nonlinear. Why not in finance? PW and Marco A. do deal with NL stuff. QuoteIt's a little bit like splines in computer graphics and video gamesThat's a bit harsh. It is more like hidden-surface removal and shading.. BTW try writing a 3d FEM engine for black-body radiation in a roundy tube:DI hear you. I talk based on my industry experience. I worked for large franchises which were favoring small-factor models with pde pricing as opposed to multi-factor model driven Monte Carlo pricing and they were actually running money on that. It's not about whether pdes are linear of non-linear in my opinion. I feel that I keep making the same point in almost every thread here. The underlying models aren't particularly meaningful from an economic perspective. Everybody is just recalibrating everything every day. I think that dealers are quite pragmatic. People have tried increasing the complexity of models to a certain degree, but when you don't get any significant benefit from it, you just leave it. It's sort of unrelated to what people work on academia or consultancies. Unless you actually have a proper portfolio and need to do something about it, how can you even approach the problem? That is why I thought that a lot of pde solvers had converged by the early 2000s. Yeah sure a lot of them are based on simple implicit methods, but they are fine. You could improve on that, but you are not going to get much out of it. Engineering is very different from finance. You are a mechanics guy, so let's see. In mechanics the recipe is to combine kinematics with a constitutive equation to get a dynamical system. That could be linear or not, be chaotic, and what not, solvable or not. But it's a pretty good recipe. In derivatives that recipe taken to a stochastic setting works as well as I claim above. Everybody recalibrates everything every day. I don't think you are going to significantly improve on what we have. But could be wrong.I was really just making the point that pdes are important, but it's a small part of the kernel of a system. It's too mature and not important enough to justify a full time role on it in the industry. I was not trying to be harsh at all. Interpolation as it is used in computer graphics is fascinating and fundamental. It's just that it occupies a similar place in a modern video game from what I understand.

As a subject for research quant finance could be one of the best: stats; maths; economics; business; behaviour. It has more ingredients than anything else I can immediately think of. But people don't have the imagination.As a practical subject it should be like engineering. And there should be fewer people doing fewer things with less money. Then it might be safer.P

QuoteOriginally posted by: CuchulainnQuoteOriginally posted by: ShihHauTanHi Daniel,QuoteIn engineering 90% of models are nonlinear. Why not in finance? PW and Marco A. do deal with NL stuff.May I ask some details? What's the full name of PW and Marco A.? Thanks!Hi Shih-Hau,PW == Paul Wilmott, the Boss Marco Avellenada. See UVM type models, for example. HJB etc.Thank you very much. In the book Nonlinear Option Pricing Julien Guyon and Pierre Henry-Labordere also discuss many complicated nonlinear models. I think these nonlinear models can have advantages, like "theoretically" can approach the market well and also allow doing calibration without using too many parameters. But it is also necessary to make the whole computation to be efficient (e.g. with GPUs) in order to make it more competitive. Otherwise recently I went to some quant seminars, and found people like using machine learning in their problem which doesn't need to solve these models.