Help me choose Master's major (ML vs Operations Research)

[plant]

I'm a CS undergraduate (3-year degree) with a minor in "systems sciences". I feel like I have made good choices thus far (feel free to correct me on that), as my current curriculum now includes for instance:

MATH:
calc 1-3
linear algebra 1-2
prob & stats
intro to optimization
numerical methods
statistical inference
prediction & time series

CS:
several programming courses
several software project courses
operating systems
databases
computer networks
intro to machine learning

MISC:
several lab / computational courses in operations research & applied math
managerial accounting and finance


My problem is choosing a Master's major (2 years) that would be the most helpful for a quant career. I'm not sure but feel like I would rather do QT / QR as opposed to dev tasks.


Option 1 - "Data Science, Machine Learning and Artificial Intelligence": from CS department with relevant(?) coursework such as:

bayesian data analysis
deep learning
reinforcement learning
federated learning
kernel methods in ML
data mining
computational methods in stochastics


Option 2 - "Systems and Operations Research: from math department with relevant(?) coursework such as:

multivariate statistical analysis
several optimization courses (dynamic / non-linear)
risk analysis
investment science
mathematical modelling
system identification
system dynamics simulation


I will have some space in elective studies to choose a limited amount of courses from the other if I choose one as a major.

In either case I will go with "Financial Engineering" as a minor, courses include:

several financial engineering courses
stochastic processes


Hopefully this wasn't too messy. I am hoping for some feedback on the courses and perhaps some advice on what major to go for.

Cheers!

 

[Andy Nguyen]

Pinging @Igna who can give you insights into QR roles and skillset needed.

 

[plant]

Pinging @Igna who can give you insights into QR roles and skillset needed.

Thank you! (new here, I assume YOU pinged them?)

 

[Andy Nguyen]

When I tag him with @, he will get a notification to the post here. You can also click on his username for a dedicated thread AMA (Ask Me Anything) which he started and has plenty of insights and wisdom.
The longer you stay here and read, the more you will learn. Welcome to QuantNet and the journey. We all have been here before.

 

[jarryds]

Well what are you more interested in taking? Grad school is extra credit, if you aren't interested in the topics, it won't do much for you. If one is calling your name more than the other, choose it.

What is will say is: I am a little hesitant to get into programs that are too specific, I think the more general the better. For example, the Data Science, Machine Learning and Artificial Intelligence would be a great option if you are planning on doing a bunch of ML in industry. Now I would be careful, I am not 100% sure of how many quant positions are working on things like this. I work on sell side and can tell you there isn't much ML work going on. I would try to track down some connections on the buy side to see what kind of ML work they are doing. I haven't spend much time on buy side, so I am not a great source of info.

I also think it is worth pointing out which departments you would be learning from. IMO, it is always smarter to learn math from the math department.

Either way, it is important for you to understand that quant is very multidisciplinary - there really isn't a wrong answer in these two choices. Make sure you are weighting your own opinions over others - you know your situtation better than we do.

 

[plant]

Well what are you more interested in taking? Grad school is extra credit, if you aren't interested in the topics, it won't do much for you. If one is calling your name more than the other, choose it.

What is will say is: I am a little hesitant to get into programs that are too specific, I think the more general the better. For example, the Data Science, Machine Learning and Artificial Intelligence would be a great option if you are planning on doing a bunch of ML in industry. Now I would be careful, I am not 100% sure of how many quant positions are working on things like this. I work on sell side and can tell you there isn't much ML work going on. I would try to track down some connections on the buy side to see what kind of ML work they are doing. I haven't spend much time on buy side, so I am not a great source of info.

I also think it is worth pointing out which departments you would be learning from. IMO, it is always smarter to learn math from the math department.

Either way, it is important for you to understand that quant is very multidisciplinary - there really isn't a wrong answer in these two choices. Make sure you are weighting your own opinions over others - you know your situtation better than we do.

Thank you for taking the time to reply. Let me provide some more info just in case:

I'm based in an EU country with 100% state-subsidized education. While this is great it also means that a Master's degree (in any field really) here is the norm, not the exception. The quant finance scene seems small here compared to other places, but I have seen some quant-like positions, e.g. in the energy sector. I would be open to work overseas for some time if the opportunity arises, but I've heard it's competitive to break in like that.

All math and cs courses are from their respective departments, probably since the schools here are comparatively small (20 000 students).

I feel like both degrees are quite general, with OR being even more so from the course descriptions (general optimization courses). The ML option would probably open more doors for "general" jobs in tech as well, which is a plus since I feel like that would be my fall back route if the quant thing doesn't work out, but at the same time I feel like the label would pigeonhole me into the quant dev role (I feel like I'm more interested in QR, QT).

Would you say that the ML master's along with the FE minor and some elective math / OR-type courses would still look good for QT/QR recruiters?

 

[Paul Lopez]

Thank you for taking the time to reply. Let me provide some more info just in case:

I'm based in an EU country with 100% state-subsidized education. While this is great it also means that a Master's degree (in any field really) here is the norm, not the exception. The quant finance scene seems small here compared to other places, but I have seen some quant-like positions, e.g. in the energy sector. I would be open to work overseas for some time if the opportunity arises, but I've heard it's competitive to break in like that.

All math and cs courses are from their respective departments, probably since the schools here are comparatively small (20 000 students).

I feel like both degrees are quite general, with OR being even more so from the course descriptions (general optimization courses). The ML option would probably open more doors for "general" jobs in tech as well, which is a plus since I feel like that would be my fall back route if the quant thing doesn't work out, but at the same time I feel like the label would pigeonhole me into the quant dev role (I feel like I'm more interested in QR, QT).

Would you say that the ML master's along with the FE minor and some elective math / OR-type courses would still look good for QT/QR recruiters?

Are you able to say which country? Places like Spain, Portugal, Germany, or Switzerland I could probably say more about. For others not so much.

 

[plant]

Are you able to say which country? Places like Spain, Portugal, Germany, or Switzerland I could probably say more about. For others not so much.

Finland.

 

[Igna]

From a buy side (quantitative hedge fund) perspective, at least at the junior level, these programs could set you up for different kinds of roles. ML for signals (event A happens and that leads to X% returns tomorrow with Y% probability) and OR for portfolio construction (given all these forecasts at various horizons how do I combine them to actually make them tradeable while maintaining the highest fidelity to my forecasts and conforming to all the constraints). Skill sets from both would make you successful in the long run. I would echo jarryds’ point that some bit of generality is nice so that you can branch out when the time comes. But in terms of recruiting, keep in mind of what your competition looks like for the QR role. For ML that might mean that some folks have done work in theoretical ML, or published a paper, or for OR they might have worked on some obscure optimization problem. So some depth and specificity that sets you apart is nice too.

 

[jarryds]

Thank you for taking the time to reply. Let me provide some more info just in case:

I'm based in an EU country with 100% state-subsidized education. While this is great it also means that a Master's degree (in any field really) here is the norm, not the exception. The quant finance scene seems small here compared to other places, but I have seen some quant-like positions, e.g. in the energy sector. I would be open to work overseas for some time if the opportunity arises, but I've heard it's competitive to break in like that.

All math and cs courses are from their respective departments, probably since the schools here are comparatively small (20 000 students).

I feel like both degrees are quite general, with OR being even more so from the course descriptions (general optimization courses). The ML option would probably open more doors for "general" jobs in tech as well, which is a plus since I feel like that would be my fall back route if the quant thing doesn't work out, but at the same time I feel like the label would pigeonhole me into the quant dev role (I feel like I'm more interested in QR, QT).

Would you say that the ML master's along with the FE minor and some elective math / OR-type courses would still look good for QT/QR recruiters?

Recruiters won't sit there and ask themselves things like "oh well why did this candidate enroll in this rigorous quantitative program and not this other rigorous quantitative program?" Enrolling in ANY quantitative program checks that box on their interview checklist.

Now from the hiring managers perspective - sure some math is more relevant than others, but I would bet that most of what you will be doing in university doesn't have a great conversion rate to what buy side is doing. You could find some QR guys on LinkedIn and see what they think about the coursework - notice I didn't say QR/QT, it honestly probably doesn't matter for QT.

We are talking about masters programs here - you will not have enough time to really distinguish yourself in some niche area of mathematics. This conversation would be way different if you were a PhD candidate and concerned you were going to pigeonhole yourself into 5 years of research on a topic nobody cares about. By the way, if QR is your goal, you should think about putting a PhD option on your radar.

Again - I don't work in QR/QT, so take this with a grain of salt. Would be great if a QR guy could step in and tell me if I am getting something wrong here.

 

[plant]

Recruiters won't sit there and ask themselves things like "oh well why did this candidate enroll in this rigorous quantitative program and not this other rigorous quantitative program?" Enrolling in ANY quantitative program checks that box on their interview checklist.

Now from the hiring managers perspective - sure some math is more relevant than others, but I would bet that most of what you will be doing in university doesn't have a great conversion rate to what buy side is doing. You could find some QR guys on LinkedIn and see what they think about the coursework - notice I didn't say QR/QT, it honestly probably doesn't matter for QT.

We are talking about masters programs here - you will not have enough time to really distinguish yourself in some niche area of mathematics. This conversation would be way different if you were a PhD candidate and concerned you were going to pigeonhole yourself into 5 years of research on a topic nobody cares about. By the way, if QR is your goal, you should think about putting a PhD option on your radar.

Again - I don't work in QR/QT, so take this with a grain of salt. Would be great if a QR guy could step in and tell me if I am getting something wrong here.

Yeah I've heard that QT interviews are more brain-teaser style questions. And also that a PhD for QR is preferred. I'm just not sure if it's worth it (if I don't like research enough) to do it for years just for the hopes of getting a job (even though it's essentially free for me) when I could gain industry experience instead. Then again, I assume that if that were the case then maybe QR roles aren't for me in the first place.

So if I decide to go with the ML option (would probably be better in general for employment / tech side as well) and still squeeze a few of the optimization courses from the OR I wouldn't be much worse off in terms of recruitment? Or alternatively go for the OR option and squeeze in a few of the most important ML courses.

I'm still not sure yet which ones of all the courses are most relevant, so I'll have to research / consult with others and then see what route could cover most of them.