Job Skills Becoming a Quant: Advice for STEM PhDs - Expert Insights

[member 428835]

Hi PF!

I'm interested in becoming a quantitative analyst, sometimes referred to as quant. If you're a quant, could you please give me your opinion of my current situation.

My PhD is aerospace engineering though my background is math heavy (just published in a mathy journal, for reference). In my work I analyze experiments (low-g fluids), theory (lubrication theory and spectral techniques), and simulations (OpenFOAM, an open-source computational package). I graduated in May and am currently postdocing. While I code frequently, I wouldn't consider myself a programmer.

I've become aware that quant is a thing, and after doing extensive research, it seems like a pretty cool fit for me. However, after reading the listings on Linked-In, it appears base requirements are PhD in STEM from a top 10 school. Fortunately I fit all requirements, but I can't help but feel I'm lacking on the programming side. Another postdoc opportunity came up that does machine-learning and robotics, and they're very interested in me. However, the school is middle rank (about 50-70 overall). Would postdocing at a school lower-tier hurt my chances at becoming a quant?

I'm not intending for this post to come across like a jerk, I'm sincerely unsure, and have asked a few quants in the field what they think, but I'd like a larger sample size. Thanks so much!

 

[andrewkirk]

I work as a quant in Sydney, Australia. In my experience, our employers have almost none of the snobbery about prestige of job applicants' universities that American employers seem to have. All they usually care about on the educational front is whether you have good marks in relevant subjects from a credible university. And credible doesn't mean 'group of eight' (our most prestigious unis). It just means not being something like Liberty University (a type of institution of which we thankfully have almost none over here).

Another apparent difference is that here they care much less about letters after your name than about what you know and what you can do. A PhD seems a pre-requisite in the US whereas here most quants don't have PhDs, but often have other more relevant higher qualifications such as actuarial, Masters of Applied Finance or CFA.

You might consider studying for a CFA. It is a three-year study program, done by correspondence mostly by people working in full-time jobs at the same time. CFA is well-respected amongst quant employers here. Being a US-based qualification, I expect it would be well-respected over there too.

A lot of quant roles are programming-heavy but not all. But I think to be a credible quant you need to be able to some simulation and pricing calcs (or logistic regression and model-building if you work in credit risk). In less complex cases those calcs can be done in Excel but generally are much better done in 'languages' like R, SAS or Python. I would strongly recommend learning at least one of those three (or C++ or C#, but I find the learning curve for those bigger) and also some basic SQL (used for data extraction from big databases).

 

[Office_Shredder]

Where are you trying to get a job at?

 

[member 428835]

Thanks for the replies. To answer the above, I know python, as well as a few other languages (no chance I could make it through the PhD without it). And I'm not sure where I want to work. Recruiters reached out, but I know so little about these roles, it's kind of intimidating looking.

 

[paralleltransport]

My impression is they just want smart people who can do math and who can code. If you fit this profile it helps a lot. You seem to fit this profile. I bet you have more programming experience than a lot of theoretical physics PhD's whose only coding experience is mathematica. I wouldn't worry about that part.

I'd worry more about how well you are at doing quick math problems. It seems those companies really like brain teasers (since those are the only things you can really ask under time constraint). I'd make sure to be ok with basic probability/combinatorial type problems.