Hi there guys,I'm hoping to start applying and interviewing for entry-level quant developer positions soon (assuming that there are currently some out there!). To practice I'm currently working through Mark Joshi's "Quant job interview questions and answers" book, as well as looking through problems that people have posted on here and elsewhere on the internet.I've assumed (possibly incorrectly) that most of the questions I've seen are for quant analyst interviews, so I was wondering: can anyone who knows about or has done (entry-level) quant dev interviews (on either side) say how they differ from (entry-level) quant analyst interviews (if they differ at all)? Is there more of a focus on programming questions? Is there any difference in the level of the financial and/or mathematical questions asked?I assume this varies a lot (depending on the position, firm and interviewer), but any insight would be helpful.Cheers!

Quant dev roles will most deffinetly have more programming questions (60%-80%?). What kind will be reflective of your background and the role more than anything else. Last one I did had questions on C++, C#, R, Matlab and Python language and public libraries specific things along with architecture, maintainability and scalability questions. The rest was on maths and finance.

To Hansi's list I would add questions regarding the performance characteristics of various algorithms and/or code. Whether you are going to be programming GPUs or not, a quant developer will be expected to write more efficient code than a quant, but won't be expected to have the same level of math or finance.

Thanks both for your input - really helpful as I've not been able to get a handle on this until now.Jim you mention programming GPUs - I've recently been wondering about this. Is this at all commonplace (and so presumably experience in this area would be valued in an entry level candidate), and if so is there a platform/set of libraries that is commonly used, e.g. CUDA?Not that I have specific CUDA experience, but if it would be valuable at entry level I would spend some time on it (I have a background in parallel hardware programming which I think would help).

In a tough job market, anything you can do to differentiate yourself (in a positive way) will help in landing a job. But I personally don't give much weight to self-taught or classroom exercise only experience. I give stronger weight to work experience where someone paid you to do something.While GPU programming experience is becoming more common, it is not usually required for an entry level position. A lot of firms have yet to make use of GPUs because they simultaneously need fast computing with lots of data (and getting lots of data to where GPUs can make use of it remains a bottleneck). So being able to write fast, efficient, numerical code will be of value to a larger set of potential employers than experience with one specific technology, particularly one which is self-taught or only experienced through classwork.

You may want to ask enough questions of the recruiter who put you forward to discern what exactly the team works on, in terms of languages, architecture, and products, then weight your preparation accordingly. It's really important to not try to go for an amorphous blob of an interview scope, but a very well defined one. Of course, it depends on how technical your recruiter is, which is not much, usually, so when you have the interview arranged you may want to send questions in writing via the recruiter, so that they are answered precisely and by the recruiting person themselves. I wouldn't say that this is a bad idea, because it shows interest and the recruiting manager psychologically invests in you by contributing their time. For quant dev you're looking at only 10% product knowledge, the rest is algorithms. Sorting is nearly always asked. You may want to make sure you have the key numerical methods down to a tee, that you know your memory management and arrays, passing by value and by ref, that you can create a class structure to solve a problem that can also be done using simple functions. Standard issues about polymorphism, inheritance, encapsulation etc should also be covered. I reiterate that it's very important that you know how to design classes instead of just simple functions and that you can quickly handwrite code that solves basic numerical problems.Some other aspects are language specific. I'd say pick the one you're best at, and focus on applications for that alone. Recruiters like to see a strong and cogent story, rather than trying to be a jack of all trades. For proxy measures of demand for skill sets, go to indeed.com and search for each language. The number of search results will tell you what you should focus on.

Thanks for the tips - really helpful, and I think I'm getting there...Cheers!

What do you mean by basic numerical problems? Numerical integration? Some numerical algebra such as LU, etc? Or more quant related FDM, Monte Carlo, binary trees, etc?

You can definitely expect questions on multi-threading, design patterns, data structures and algorithms, Boost library, STL library. Then a few Math questions. Sometimes they give a coding exercise just to make sure that book knowledge alone is not enough to get the job.

I agree coding exercisesparallel execution lock mutex don't forget brainteasers maybe both related to computer science algorithms and probability ones.