For those who landed as a quant with only undergrad degree

Dorris's picture
Rank: Senior Chimp | 29

How strong is your math background? I have taken real analysis 1, 2, machine learnig, bayesian inference. I am taking measure theory next term. I am worried thst i'm lacking so much though. Differential equations, stochatic calculus, forecasting and the list goes on. I know it may depend on the area of quant finance, what are the must-know topics?

Comments (29)

Dec 7, 2017

Landed an HF quant role with pretty much just calc 1-3, linear algebra, differential equations. My major is a very quant-heavy hard science, though, and I use a lot of applied math in my major-specific courses.

Dec 12, 2017

Were you able to land a HF quant role straight out of undergrad? From what I've read I need to "pay the due" while working in IBD or S&T for a couple of years.

I went to a (semi-) target school but I'm not so sure if I can land on a job straight out of undergrad. I'm currently working in tech.

Dec 13, 2017

Yes, straight out of undergrad.

Best Response
Dec 7, 2017

Your background is looking good mathematically speaking. I would make sure to load up on more stats classes. Time series, theoretical stats would eb good. Differential equations is mostly useless. Stochastic calculus is useful for the sell side but not as much on the buyside i don't think. Maybe for specific hedge funds that deal with more exotic products or market makers.

Programming is important. Become an expert in Python (numpy, pandas, scipy, TensorFlow). Become at least very proficient in R. Being able to code in C++ or Java is also often useful if you're going to be working closer to the production side of things.

Anyway, yeah, your math background is fine. You aren't lacking. The important thing is that you are able to comfortably read and understand research papers. Take a look at some stats or machine learning papers that have come out from the main journals recently. Can you grasp them pretty quickly, understand the conclusions and think of some potential applications? If so, you're doing fine. Usually the applicable papers you would read don't really require real analysis or measure theory. Taking those classes just allows you to more quickly understand the math because you're used to thinking in that way. But being able to understand or do complicated proofs isn't really necessary

    • 4
Dec 9, 2017
DeepLearning:

Your background is looking good mathematically speaking. I would make sure to load up on more stats classes. Time series, theoretical stats would eb good. Differential equations is mostly useless. Stochastic calculus is useful for the sell side but not as much on the buyside i don't think. Maybe for specific hedge funds that deal with more exotic products or market makers.

Programming is important. Become an expert in Python (numpy, pandas, scipy, TensorFlow). Become at least very proficient in R. Being able to code in C++ or Java is also often useful if you're going to be working closer to the production side of things.

Anyway, yeah, your math background is fine. You aren't lacking. The important thing is that you are able to comfortably read and understand research papers. Take a look at some stats or machine learning papers that have come out from the main journals recently. Can you grasp them pretty quickly, understand the conclusions and think of some potential applications? If so, you're doing fine. Usually the applicable papers you would read don't really require real analysis or measure theory. Taking those classes just allows you to more quickly understand the math because you're used to thinking in that way. But being able to understand or do complicated proofs isn't really necessary

Where do you usually find the latest research papers?

Snootchie Bootchies

Learn More

814 questions across 165 hedge funds. 10+ Sample Pitches (Short and Long) with Template Files. The WSO Hedge Fund Interview Prep Course has everything you'll ever need to land the most coveted jobs on the buyside. Learn more.

Dec 11, 2017

Kind of everywhere. SSRN for more finance specific papers. I follow a lot of academics and professionals on LinkedIn and Twitter who will post or comment about new research. Annals of statistics is also pretty good. But generally I'll just read what some of my favorite experts in stats, machine learning, quant finance are posting or reading about

    • 1
Dec 12, 2017

Work in a different field but have recently begun learning R. Curious what, if any resources, you'd recommend to help augment the learning process?

Dec 12, 2017

Can't think of any specific resources for R at the moment. I learned python-pandas before I learned R and the syntax is very similar. Mostly you just need to learn how to use the different apply functions in R and the other stuff is done for you. No silver bullet resource for R or any other programming language unfortunately. Just takes time and effort.

Do one difficult project in a language and you'll learn it through blood, sweat and tears. After that it's easy.

    • 2
Dec 12, 2017

Master one of data.table or dplyr, and you will be unstoppable. R is but a toy language without one of those libraries.

Dec 12, 2017

Agree. I'll chime in with two addenda: first, candidates are normally overqualified on the math side and extremely underqualified on the programming side. In my experience, the programming side is much more important than the math side, even for researchers. The math classes are important for an entirely different reason: besides teaching you how to think, they also give you the expertise to know when simply taking the arithmetic mean is the correct approach.

Dec 12, 2017

This is great! I'm surprised to hear that DEs are mostly useless. I thought (stochastic) PDEs are bread and butter of quants.

As for programming, I've taken classes in Python, C++, R, Data structures and OOP. I'd say I'm pretty comfortable with all, unless the firm requires ultra-low latency system architecture stuff (!!! I don't even know if this makes sense)

Well, I read some research papers on arXiv and mostly struggled with them because a lot of them assumes knowledge of stochastic calculus. A lot of Ito's calculus and martingales show up and I'm usually left clueless after leading one or two pages.

Taking measure theory is more of a personal interest, but it is provides elegant approaches to problems in probability theory from what I heard so I think I'll enjoy it very much. I have one more room, and currently debating between financial mathematics, differential equations, nonparemetric regression, or even differential geometry (!!!)

I guess my pure math knowledge is somewhat passable, but lacking a bit of statistics. I haven't done any forecasting or time series and I don't know if I can pick it up while on the job so I can take more interesting courses.

Dec 11, 2017

I think people are overdoing it. All this quant shows mathematical aptitude/ ability/ interest, but basic programming and some quant skills are all you actually need. If you are applying to a quant-heavy place, just pick a couple of topics and get good... and not real analysis, topology, number theory, etc... something more applied. I have seen people mention Dastidar's investments book on the forum; that has intermediate level material for both quant and discretionary folks. Starting with that is a decent choice.

    • 2
Dec 11, 2017

^This, I did a SA stint for a Quant firm and went through plenty HF interviews for FT. Just take one or two stat classes and put some shit like "Applied Regression, Time Series and Probability Analysis" on your resume and you'll do just fine. There are a few funds that ask brain teasers where you might need to apply a simple probability or stat concept but it's not as difficult as it's hyped up to be.

Dec 12, 2017

This is kind of terrible advice, I'm sorry. Even if this is the case at some quant funds now, it won't be in the near future. With the rise of machine learning, alternative data and use of non-linear models in general, the expectations are going to be higher. At more fundamental funds, I'm sure only more basic stats is necessary but if you actually want to work on the research side of a quant shop, you need to have a strong math background.

Dec 13, 2017

But 9 out of 10 job postings I've seen, a master's degree is a minimum so I feel like I'm a little behind. I understand there are separate career tracks for quant with just a bachelor's degree but I haven't seen a lot of jobs like this.

Well I don't see topology or number theory can be directly useful but aren't real analysis or differential geometry useful for probability theory or PDE theory? I thought quants are physics or maths PhD who live and breathe differential equations and just use them all day.

I'm reading Quantitative Finance for dummies by Steve Bell, and the amount of material is, as I worried, quite large. It mentions the basics from stochastic calculus to PCA. I doubt even graduate programs that cater toward quantitative finance cover everything. On the other hand, it was a very cheap book...

Dec 13, 2017

Quantitative finance is not all encompassing. Exotic derivatives people are quants. There are commodities people who are quants. There are options pricing people who are quants. There are HFT people who are quants. There are long only investors who are quants. There are fixed income people who are quants.

Stock brokers and investment bankers might both use Excel but that doesn't mean their jobs have all that much in common. Similarly, while quants across different areas of finance may both use math, they use different types of math.

Stochastic Calculus is very important to quants on the sell side who are pricing options and other derivatives. If you want to deal with options and other complex financial products, stochastic calculus is important. This is not my area of expertise. I actually know very little stochastic calculus.

Real analysis really just provides the language and structure for thinking, reading and writing mathematically. Knowing real analysis gives you a strong toolbox for reading mathematically heavy text. It isn't particularly useful itself. Crazy advanced probability theory is not often used. More commonly, it is a combination of upper level undergrad/lower level graduate probability and statistics. However, what makes the 'real world' and graduate school much more difficult than undergrad math classes is that you have to understand and use a combination of concepts from the undergrad level. They aren't silo'd off in separate classes.

Differential equations are used in stochastic calculus. I don't know much stochastic calculus so I can't speak to it. But generally, differential equations are used to model more deterministic systems. Obviously, financial markets are not deterministic systems.

In the quantitative equity world where I live, it all comes down to finding and/or generating useful/predictive signals and then coming up with a methodology for using those signals to rank/select stocks to long or short. For the most part, methods in this space use upper level undergrad/lowever level graduate probability, stats, time series analysis and more recently machine learning techniques.

    • 3
Dec 13, 2017

Depends heavily on how 'quant' the quant firm is- I think trying to paint a picture with too broad strokes here isn't useful.

If you're talking Point 72/AQR, you for sure do not need to know machine learning for the interview (unless the role is machine-learning specific). Its not an efficient to filter out people using a topic that few people study, and that you have no chance of knowing unless you study.

Instead, to select for people who can program (and are mathematical/logical), expect basic programming questions, brainteasers, and questions about things on your resume that are quant. So if you have machine learning there, you may get asked about it.

Dec 14, 2017

yeah ur right that machine learning isn't a requirement by any means. However, it is much much easier to differentiate yourself at the undergraduate level by knowing it since that was the original topic of the post.

AQR is only recently getting into machine learning from my understanding.

Dec 14, 2017
Comment