I’ve found a lot of helpful information browsing this forum for the last few days. I am left with some questions about finding a job in quantitative finance. I have a PhD in physics (theory) from a top school and three years of postdoc experience doing simulations. I am a very experienced numerical programmer, but in FORTRAN. I passed the CFA level 1 exam in December. My 6 month plan is to switch over a small numerical project to C++ (learning c++ in the process) pass the CFA level 2 test in June, do self study in excel & VBA, and read several finance books (Hull, Neftci, Wilmott, etc.). I’m also considering the PRM or FRM exam. Then, I will quit my physics job and go look for a quant job.So my questions: if I were to quit physics today and look for a job, how difficult would it be and what kind of starting compensation would be offered. Would the 6 month plan outlined above very significantly increase chances of finding a job or add to the starting compensation? Any other advice on how much preparation should be done before quitting physics and committing to a full time job search? Thanks in advance.

Hi lifoThis is a pretty common situation, and I am happy to let you know that it is possible to make the switch, but there is a LOT of work to do.My #1 piece of advice would be to focus. It's great to be able to tick all the boxes on the headhunters form, but a deep knowledge of one or two areas is much better than a broad knowledge of this big thing called "deriviatives".You need to spend time looking around the finance industry and try and get a feel for what areas are "hot" right now, and where you might have a competitive advantage. Have you come across anything in finance that interests you? For example, if you thought credit derivatives were an interesting area and had written your thesis on predicting earthquakes, then you should be able to nail an interview after only 3-6 months of study and making contacts.Understand your strengths, adapt them to the problem, and work hard - you will get there.

QuoteOriginally posted by: lifoI’ve found a lot of helpful information browsing this forum for the last few days. I am left with some questions about finding a job in quantitative finance. I have a PhD in physics (theory) from a top school and three years of postdoc experience doing simulations. I am a very experienced numerical programmer, but in FORTRAN. I passed the CFA level 1 exam in December. My 6 month plan is to switch over a small numerical project to C++ (learning c++ in the process) pass the CFA level 2 test in June, do self study in excel & VBA, and read several finance books (Hull, Neftci, Wilmott, etc.). I’m also considering the PRM or FRM exam. Then, I will quit my physics job and go look for a quant job.So my questions: if I were to quit physics today and look for a job, how difficult would it be and what kind of starting compensation would be offered. Would the 6 month plan outlined above very significantly increase chances of finding a job or add to the starting compensation? Any other advice on how much preparation should be done before quitting physics and committing to a full time job search? Thanks in advance.it's pretty unusual to do things like the CFA exam before applying for quant jobs. of course that might help your cv sitck out... The main things I would concentrate on are getting a good knowledge of derivatives pricing, and the practice of implementing models in C++.If you are any good it shouldn't be that hard to get a job but that largely depends on the state of the job market at the time you are applying. Learning more before you start looking is unlikely to get you a better starting salary but will help you get a job.It's unwise to focus too much on the starting salary; it will be substantially more than a physics postdoc gets and will go up.

Thanks, EnergyQuant and mj, for the advice. Looking back, the CFA material is much better preparation for financial statement analysis than quantitative work. But it is good to get the big picture and learn what’s interesting. (At least that’s what I tell myself after having bought all the books and paid all the fees). It sounds like I should do some preparation before jumping in—definitely start programming in C++ and find something to focus on. Is it fair to say that Hull’s book has a little about everything, or is there a better overview book? (I found the “paths to enlightenment” thread in the book forum, but there’s no way I can read all those books in 6 months). Any advice about where to find out what’s becoming hot now, or hot in 6 months? These forums, a magazine…?

you could do worse than to work through the computer projects in my book. (in my unbiased opinion)personally i am not a great fan of hull -- the target audience is business students not maths and physics phds.

Actually if you look under careers, sometimes orgtel just look for Phd in Physics, astro physics w/o prior experiencesso I guess actually you are more than able to switch to finance. From the forums posted by Orgtel

I also came from physics background and now I’m doing basic finance and C++ reading. I think I have some idea as to which fin. math books will help me to become employable in the nearest future, but what level of C++ knowledge is expected from a person with my background is still a mystery for me. I’ve never done any programming but worked through the 1st volume of Eckel’s book (just played with the code, didn’t do any exercises). Of course I understand that the best way to learn a language is to read and write as much code as possible but I’m under severe time constraints, moreover there’s no guarantee that some project I can do for myself will actually help me during the interview. I mean of course I understand that there is a great variety of positions with widely varying level of expected programming experience, but for example ads like:“…extensive C++ industrial experience or relevant coursework are required…” really confuse me, I mean how can you compare real ‘work experience’ with a ‘coursework’?...So my question is what kind of quant positions I should look for if I have rather strong math background (PhD in fundamental physics) when I will have basic financial math knowledge, but no programming experience. I mean I will be eager to obtain this experience working on real projects but not on some imaginary ones. Also except mj’s book (which is still unavailable from amazon in US) are there any books which address directly C++ application to finance (of course I’ll try to search Books forum as well), and lastly is it enough to have only basic C++ knowledge, but no Excel, VBA, etc. to start a job hunt?

the list of computer projects in my first book which is already out. re C++ books, there are 3 coming out in the near future including mine but none are yet available. The brooks book seems to be singularly unpopular. My team has taken people with no C++ experience in the past but only if they had extensive programming experience in another language. I can't imagine any UK or US quant team taking someone who can't program.

Learning C++ will make you more marketable than Fortran.Once you've worked out how to put the appropriate wrapper around your C++ to make it compile and do input/output, you won't find much difficulty doing straight numerical stuff after Fortran. This is both good and bad.The problem is that anything you say you know, can be asked in an interview. C++ has subtleties and complexities that even people who program 50 hours a week for three years get wrong. Thus your choice is to say in effect , "I've done Fortran, and am smart enough to master C++". In which case you will be asked questions on Maths/Physics, and possibly Fortran. Thus you get asked about your home subjects.If you say you've done C++, you may get asked how to call a pure virtual function or the role of constructors in STL.On average, recruiters for entry positions are looking for potential with some vector towards their current needs. Thus it is frequently better to be excellent in something useless to them, like Fortran than be mediocre in a relevant skill.

Hi, I am having a similar experience. I have a PhD in physics and a couple of years of experience postdoc-ing. I recently decided to give finance a go and I was quite serious about getting up to speed, that means that I am spending a lot of time reading and doing problems. I have been fairly successful in securing some interviews with all sorts of places and my impressions so far - I am nowhere near to getting an offer - are as follows. OK, it is a weird world. I think that someone in this thread mentionned that it is hard workand I confirm this. Personally I spent a lot of time reading Hull and some time Baxter&Rennie. I did the problems in the Crack book, then also read most of Crack's new book on basic BS. Meanwhile, I am down to consilting a mixture of Baxter&Rennie, Bjork, Wilmott's student book,random probability texts, Lamberton&Lapeyre, various problem sets I found on the web.etc..... Ah yes, and I have been extending my knowledge of C++ beyond the basic subset I am using in my daily research work.What surprised me most, was that I had been told that one could away with a decent knowledgeof Hull, some physics-type stochastic calculus and maybe one better treatment, let's say B&R. I believe that under luck it is possible to score a job with that. However, I have been asked all sorts of stuff. From basic finance questions (whose answers must become second nature), to applied maths questions (level of typical methods course, such as Cambridge tripos course methods 1B), to basic probability questions, to stuff I would term more advanced, such as demonstrating the martingale representation theorem, etc.... Totally weird, it is difficult to know what to know and what to prepare. I argue it is fun to do that preparation, but it is positively more than bedtime reading you need to do, I think. I have so far not been able to find out, to what extent a good grasp of risk-neutral valuation in a large sense is required and actually used in banking. I think that getting a decent understanding of that is a challenge for most physicists as this forces to take quite a detour into probability, stochastic processes, etc..... Our basic probability education is just too pathetic and hopefully future statistical mechanics courses will take care of this. unkpath

>to take quite a detour into probability, stochastic processes, etc.Add to this some control theory, dynamic programming, fractals, chaos and genetic algo'sand you"ll be in good shapeMy first Q in my first interview was from optimal stopping time theory

My study plan is very similar to unkpath’s. I really enjoyed reading Baxter & Rennie (I worked through it in 4 days), reading Hull is rather depressing but since it’s necessary I put up with it.Unkpath, how do you know you’re “nowhere near getting an offer”? Did you screw up some interviews completely? May be you can share your experience with regards to the questions in some more detail (PM me if you prefer). In particular, programming questions are the mystery for me, I’ve heard that sometimes they ask pretty simple questions, but sometimes ask to find subtle bugs in the code or to write an option pricing program, it’s hard for me to imagine how could it be done in such a limited time.mj, thanks for your opinion, but the phrase 'extensive programming experience' still confuses me, what does it mean for a person who never had a job? can it be tested in half an hour(?) during the interview? Is taking ‘advanced’ level C++ course ‘extensive’ enough? I’m asking because I was going to learn basic numerical simulations techniques (to be able to answer simple questions) and then put on my CV that I’ve done all that during my PhD years. This idea occurred to me after talking to a friend of mine who does some kind of numerical simulations for solid state research. Those simulations are running on his computer non-stop for 3 or 4 years, but when I asked him he told me that during all this time he has written just a couple of programs and after that just modified them slightly for each ‘new’ project. In principle I could have been doing the same, right

There has been a lot of good info here. Thanks to everyone. I have some more questions about applications/interviews:I am a US citizen. Does that mean it will be much easier for me find a job in New York than outside the US? Would London, Hong Kong, etc. be difficult for someone without the right passport?My postdoc position is in Asia. Is it worth applying for jobs in New York or London while I am on a different continent and unable to interview in person?I sometimes get the opportunity for a couple of days stopover in Hong Kong, London, or New York on the way to conferences. Is it worthwhile to apply for positions when I will only be around for a couple of days? Or is it better to make a clean break from physics, move to where the jobs are, and be a full time job hunter until employed?

A somewhat related question... what are the bottomlines of these C/C++ programming skills ?

I have just a Bachelors in applied physics and am thinking of doing a Masters in Math Fin somewhere down the line. I think thats the best prep that can be. Also considering taking prob courses to get to speed for the Masters.