• C++ Programming for Financial Engineering
    Highly recommended by thousands of MFE students. Covers essential C++ topics with applications to financial engineering. Learn more Join!
    Python for Finance with Intro to Data Science
    Gain practical understanding of Python to read, understand, and write professional Python code for your first day on the job. Learn more Join!
    An Intuition-Based Options Primer for FE
    Ideal for entry level positions interviews and graduate studies, specializing in options trading arbitrage and options valuation models. Learn more Join!

How to prepare for a quant job after doing a Ph.D. in physics (Quantum Information) ?

Joined
9/9/20
Messages
2
Points
11
Dear all,

I am an international student who is going to complete a PhD in physics in Australia in about six months (specifically, I worked on quantum communication protocols). Before my PhD, I studied Electronics and Communication Engineering in my undergrad. I am considering to shift to industry, possibly in Hong Kong, Singapore or Australia for the next 4-5 years and eventually get settled in India. I am looking for jobs that involve a fair bit of mathematics. With that in mind, a little bit of research introduced me to quantitative finance. I am writing down some aspects about myself as objectively as possible.

Particular mathematics that I have used in my research:
Linear algebra, probability, information theory, signal processing.

Exposure to coding:
  • Matlab: I have a good exposure to Matlab, for example, I have a paper where I modelled atmospheric turbulence. I have also created several optical signal processing in Matlab. I have used a convex optimisation package, like CVX.
  • Python: Although I am not as fluent in it as in Matlab. Mainly it boils down to do a quick google search for the exact Matlab equivalent command that I am looking for, and after that I can fairly easily implement the algorithms. Also I have a good of exposure to the libraries like sci-kit learn.
  • C: In my undergrad, I have done basic programming algorithms like sorting, printing patterns, data structure. However, I have not prepared particularly for any coding interview.

Exposure to machine learning and deep learning:
  • I have done a deep learning workshop from Spaceport AI that involved 3 Kaggle projects. There were projects like twitter sentiment analysis or recognising cats and dogs.
  • Currently, I am working on a physics project that involves machine learning, the work should appear on the arXiv soon.
  • I am currently doing a Udemy certification course.
  • However, I haven't taken a machine learning as a core course, I know a bit of linear algebraic model behind neural network or the principle behind gradient descent algorithms but if someone asks me to solve an exercise problem from Christopher Bishop, I will get stumbled for a while. The point is, if it is worth investing time to get well versed in the underlying math I can do that.
Soft skills:
I am an introverted person, however, my Ph.D. gave me several opportunities to improve my communication and presentation skills. Based on my personality, I would like to be the guy who works on mathematical and statistical modelling and decode the state of the art trading papers.

Weaknesses I can think of:
  • I don't have a finance background.
  • I haven't worked on SQL, however unlike data science jobs, I have seen that most quant ads don't have this requirements.
  • I need to prepare myself with coding interview questions.

Questions:

1) How much chance do I have if I want to land a quant job immediately?
2) Assuming the chances are low for 1), what do I need to work on if I want to land a job before, say June 2021? Realistically, I will be busy in writing a few papers and after that I will write my thesis. So I can dedicate 2-3 months for job preparation.
3) If I do a post doc, in a field say quantum machine learning, then how will it affect my job prospect? (There are articles about doing finance using quantum computers, but I am not sure how much my degree will boost my credibility.)
4) How secure and stable is the job market, now given the COVID situation?
5) Given this pandemic situation, I don't have much opportunity to do networking. Can you guys suggest strategies or opportunities to do networking?


Any advice would be highly appreciated!
 
Hi mate, I can talk a bit about my experience of switching to quantitative finance. Interesting enough, I spent two years on researching QKD for my Master's degree but went on to earn a PhD in semiconductor photonics.

To be honest, skills earned during quantum information research are not very applicable to finance. Hard quant positions require good coding skills, effective and efficient, and strong optimization/statistical/probabilistic/stochastic calculus knowledge. That's my impression. Many friends of mine who switched to finance had to really brush up coding skills and cram relevant knowledge + finance nomenclatures.

Therefore, it's good that you have been involved in machine learning projects and taking online courses. It would help open doors to data science positions in general. You background in MATLAB, Python and C++ seems okay too. If you already have experiences in some packages E.G CVX, sci-kit learn, you can strengthen/expand your understanding of the theoretical parts too. While programming in MATLAB, Python and C++ can be more of an art, SQL is really straightforward so you don't need to worry about that.

Forget about quantum machine learning/quantum cloud computing research experience. Many of my friends, myself included, would only LMAO. No offense to anyone's devotion. If you would like to be a professor, that's a different story. But I assume when you look to quant finance for career, you are realistic. Under current conditions, job markets seem much more turbulent but the majority in the field seems to be doing fine.

You haven't name any specific groups, departments or fields. To me, you look really good to get positions in market risk developers, credit risk modelers, and generally data scientists in finance. Buy side quantitative researchers, or desk quants may be a bit of stretch due to fierce competition, but not impossible for junior roles. Sometimes luck comes into play too. At the very least, HR/hiring managers seem to trust PhD grads and would extend interview opportunities, at which point it's up to you to prove yourself. Be focused, think it over, and put together an integrated set of skills/knowledge for your interviewers. You really just need a place to start.
 
While programming in MATLAB, Python and C++ can be more of an art

Not exactly.
The scientific approach is to learn C++ as it is the basis for everything
And program like crazy.

In my undergrad, I have done basic programming algorithms like sorting, printing patterns, data structure. However, I have not prepared particularly for any coding interview.

Industry is much different. It can be/will be a culture shock.
 
You should start applying for internships now. The annual hiring cycles are about to start around this time of the year, and you'd then be looking to start work early summer. So ideally you'd do a 6 month internship and manage to convert it to a full time position. Now covid may have cut fulltime hiring in some places, so converting to a fulltime position may not be possible, but at the very least you'll then have a much improved CV, and a better taste of the industry and hopefully a better idea of what kind of a role you would like.

Given you seem to have no programming experience (one-man bash/matlab/python scripting doesn't count as such), this is the area that you should concentrate on immediately, for a strong background in a systems language like C++ will immediately make you a more employable candidate. You will also want to spend a couple of months practicing interview questions in general (not just coding), and so the time to start really is now.

Some banks (at least GS, JPM) have teams doing quantum computing (and have published papers on the subject, too). You could try and find the relevant names and reach out to them on LinkedIn or whatever and see if they might have openings. That is, if you want to stay on the academic side of things rather than finance per se.

As for job market, with a bit of Googling, you will see that some of the firms that advertise their jobs openly are still hiring, and by the looks of it, some are even expanding. There's always demand for quants with a couple of years of experience on their CVs.
 
Thanks, ZFL, Daniel, and KillingField for the input.
'Interesting enough, I spent two years on researching QKD for my Master's degree but went on to earn a PhD in semiconductor photonics.'
ZFL, We seem to have traversed the same paths. :)

'You haven't name any specific groups, departments or fields. ':

Actually I am not sure about any particular fields, I have briefly read some articles on risk management, algorithmic trading, portfolio management. There is a chapter on Thomas and Covers' 'elements of information theory', called 'information theory and portfolio theory'. I have read the chapter a bit. However, I don't have much insight into these fields, and that's why I am asking the people who are already in the game. Particularly I would be looking for a field that has well-balanced mathematics and coding (with emphasis on mathematics).

'... for a strong background in a systems language like C++ will immediately make you a more employable candidate.'

Question:
Can you guys tell me how do I demonstrate a strong background? I can certainly prepare for the interview questions, but to get an interview offer in the first place I would need a portfolio. If I work with Daniel Duffy's 'Introduction to C++ for financial engineers', solve some exercise and projects, and put them in GitHub, will it help?

'Now covid may have cut fulltime hiring in some places, so converting to a fulltime position may not be possible,'

Thanks for the suggestion to do internships. However, the above quote is exactly my concern. Under normal circumstances, I would have invested the time to brush up the necessary skills. The thing is, I have a fairly good chance to start a Post Doc as I have a good network. If I want to make the position secure, I could start a collaborative project in my Ph.D. while I am writing my thesis and continue the project on my postdoc. It will secure my finance for at least two years. Changing the field right now seems to be a bit risky to me. The other plan would be to be enrolled in the post doc and meanwhile, keep building my profile. But who am I to be so sure? :) That's why I am asking the people who have the pulse of the market.

Question:
In that context, does having a post-doc degree reduce chance in the market?

'Some banks (at least GS, JPM) have teams doing quantum computing (and have published papers on the subject, too). '
This is another great suggestion. I looked up, IBM is another good option. Indeed this is a relevant paper. I will keep that in mind.

Question:

Can you guys tell me whether having a paper like this increases chance in the mainstream job market? Here the authors are at least showing some financial background.

Thanks for your suggestions again.

 
Actually I am not sure about any particular fields, I have briefly read some articles on risk management, algorithmic trading, portfolio management. There is a chapter on Thomas and Covers' 'elements of information theory', called 'information theory and portfolio theory'. I have read the chapter a bit. However, I don't have much insight into these fields, and that's why I am asking the people who are already in the game. Particularly I would be looking for a field that has well-balanced mathematics and coding (with emphasis on mathematics).
This is something you'll need to figure out yourself. There's plenty of material out there, and once you have some ideas of what might be of interest to you, you can then ask more targeted questions. The hiring processes and the wanted skillsets can be quite different buy side vs sell side (and of course say HFT is different from asset management, too etc).

The term "quant" is also next to meaningless and almost everyone is attaching it to their job titles because it's sexy, even when there's no math involved. If you intend to return to India at some point, several major banks and funds have low cost centres there, with quants, too (though their role will be more along the lines of supporting other quants rather than a desk directly). So if you have a few years' experience at a major financial hub, I would imagine you to have no problems at all finding employment in India, and probably getting promoted on the corporate ladder, too, should you want to make the move back. I don't know how many quants the domestic institutions themselves there employ.

'... for a strong background in a systems language like C++ will immediately make you a more employable candidate.'

Question:
Can you guys tell me how do I demonstrate a strong background? I can certainly prepare for the interview questions, but to get an interview offer in the first place I would need a portfolio. If I work with Daniel Duffy's 'Introduction to C++ for financial engineers', solve some exercise and projects, and put them in GitHub, will it help?
You put C++ on your CV, and be prepared to answer questions is the jist of it (e.g. at the level of "Effective C++"). One question will be which projects you've used C++ on, and the larger the project, the more impressive it will be. Ideally it would not be a side project, but something that you did during your PhD. Nobody will be interested in you showing mastery of syntax, which is what small exercises are about. If you want a "portfolio", either make something that's useful, or contribute to some open source projects (demonstrating that you know how to work with large codebases and design code that others can read). That said, I don't think a portfolio is something that is expected of you.

'Now covid may have cut fulltime hiring in some places, so converting to a fulltime position may not be possible,'

Thanks for the suggestion to do internships. However, the above quote is exactly my concern. Under normal circumstances, I would have invested the time to brush up the necessary skills. The thing is, I have a fairly good chance to start a Post Doc as I have a good network. If I want to make the position secure, I could start a collaborative project in my Ph.D. while I am writing my thesis and continue the project on my postdoc. It will secure my finance for at least two years. Changing the field right now seems to be a bit risky to me. The other plan would be to be enrolled in the post doc and meanwhile, keep building my profile. But who am I to be so sure? :) That's why I am asking the people who have the pulse of the market.

Question:
In that context, does having a post-doc degree reduce chance in the market?
I don't think one postdoc hurts your chances per se, but it doesn't help either. Joining a firm after a postdoc, your boss may well be a few years younger than you - academic experience usually does not count as experience. Ideally you would still do an internship before the postdoc (something that I've seen some people do), so your CV would demonstrate earlier interest in finance.

I understand that you see it as risky to leave the comfort of what you've been doing for the past several years, but if you don't have the intention of staying in academia for the long haul, it is also risky to take poorly paid, insecure postdoc positions that don't advance you towards your long term goals. Of course you could take the postdoc and try applying for jobs, too. You can always renege or quit half-way from the postdoc. These are big life decisions and I think it's only right to be a bit selfish here.

'Some banks (at least GS, JPM) have teams doing quantum computing (and have published papers on the subject, too). '
This is another great suggestion. I looked up, IBM is another good option. Indeed this is a relevant paper. I will keep that in mind.

Question:

Can you guys tell me whether having a paper like this increases chance in the mainstream job market? Here the authors are at least showing some financial background.

Thanks for your suggestions again.
What does mainstream mean? I'm sure you'd get an interview with the group that published that paper if you'd done something similar during your PhD. There's an increasing interest in quantum computing, though I don't personally see it revolutionising quantitative finance (there are plenty of shortcomings in the models themselves that don't have to do with lack or compute etc).
 
Given you seem to have no programming experience (one-man bash/matlab/python scripting doesn't count as such)
yeah this is my current problem. don’t get me wrong i enjoy just doing regressions all day but this isn’t going to prepare me for the top quant shops which increasingly want to see legit software engineering experience. what would you say i should supplement my day to day work with to gain this skill (as much as is possible, from the outside of an industrial production environment)?..leetcode to pass the interviews?
 
yeah this is my current problem. don’t get me wrong i enjoy just doing regressions all day but this isn’t going to prepare me for the top quant shops which increasingly want to see legit software engineering experience. what would you say i should supplement my day to day work with to gain this skill (as much as is possible, from the outside of an industrial production environment)?..leetcode to pass the interviews?
You do have a good point here: Rather than learning programming you could use your time to target the kind of questions that are likely to come up in the algorithms part of an interview, which for a SWE-type role in particular will be the main material. Leetcode and friends are good practice for this. This of course is not programming per se, and you'll not learn about organizing code or best practices (e.g. toolchains or patterns) and are unlikely to come across anything deeper about a specific language (e.g. what is a vtable or how to implement std::move). Actual programming questions (like the examples before) are more difficult to ask in an interview and so there's often less focus on them, but as they usually do come up in short-form, I would reiterate my recommendation to read through something at the level of "Effective C++" in your target language, anyway (or if you had more time, get involved in some open source projects of interest to you).
 
You do have a good point here: Rather than learning programming you could use your time to target the kind of questions that are likely to come up in the algorithms part of an interview, which for a SWE-type role in particular will be the main material. Leetcode and friends are good practice for this. This of course is not programming per se, and you'll not learn about organizing code or best practices (e.g. toolchains or patterns) and are unlikely to come across anything deeper about a specific language (e.g. what is a vtable or how to implement std::move). Actual programming questions (like the examples before) are more difficult to ask in an interview and so there's often less focus on them, but as they usually do come up in short-form, I would reiterate my recommendation to read through something at the level of "Effective C++" in your target language, anyway (or if you had more time, get involved in some open source projects of interest to you).
Hm...I have had to customize some sklearn transformers for my own purposes, something I can see being generalized and released back into the open source ecosystem.

Do you have a recommendation for a book like 'Effective C++' but for Python? I know this forum thinks Python is a bad language to learn legit programming practices in, but idk I see tons of roles nowadays for Python developers. And pretty much all QR roles desire that if not making it a requirement, so it is a language that is not just used for pandas scripting.
 
Hm...I have had to customize some sklearn transformers for my own purposes, something I can see being generalized and released back into the open source ecosystem.

Do you have a recommendation for a book like 'Effective C++' but for Python? I know this forum thinks Python is a bad language to learn legit programming practices in, but idk I see tons of roles nowadays for Python developers. And pretty much all QR roles desire that if not making it a requirement, so it is a language that is not just used for pandas scripting.
Python isnt a bad language to learn programming per se. If you are taught bad programming habits, it could be in any language.
 
Back
Top