I would like advice on how to become a quant

I am a data-engineer turned software engineer. My intention is to become a quant developer or possibly a quant researcher (although this might not be right for me). I would like to ask successful quants for advice and resources to aid me in my journey.

- Development projects.
- Books & courses.
- Relevant mathematics to study.
- Software to learn.
- Relevant machine learning.
- Most profitable area of finance technology.

To give some context, I am fluent in Python, SQL, VBA and written some R, (haven't written it in years) and I have recently developed my own implementations of descriptive statistics (basics), k-means clustering and now linear and logistic regression. I have weak mathematics background which I am improving, and ideally would like to greatly accelerate my learning. I didn't finish my degree and feel that I am now past doing one maybe a masters in machine learning and AI once I feel I am sufficiently competent but unsure.

Finally, I really appreciate any advice. :):thumbsup:
 
C/C++ is fundamental, you need a little bit more than being "fluent"
Do you mean:
- asymptotic analysis
- algorithms and data structures
- knowledge of existing packages and tools
- testing
- integration between api etc

I am familiar with asymptotic analysis, I know quite a few algorithms (weak in dynamic programing but could pick it up easily enough) and have average library knowledge, I also test my work and have integrated software.

I could learn C or C++, which would you recommend and for which purpose? My understanding was that for modelling Python is often used and for algo trading C++. There are many quant jobs asking primarily for Python in their advertisements.
 
Can you order a coffee in Python.
1611491899805.png
 

Daniel Duffy

C++ author, trainer
"Fluent" could mean anything. Can you program?

For C++ it means being able to write programs, from scratch if need be.
 
Last edited:
"Fluent" could mean anything. Can you program?

For C++ it means being able to write programs, from scratch if need be.
If I have not implemented something before i do so from first principles. Taking the machine learning examples listed above, i wrote my own little libraries and compared the results to that of sci-kit-learn. I can use books, blogs and wikis to piece and understanding of the mathematics together. So certainly yes in Python. I suspect i could put something together in C to although i wouldn't expect to make best use of the language features and might make mistakes.
 

Daniel Duffy

C++ author, trainer
If I have not implemented something before i do so from first principles. Taking the machine learning examples listed above, i wrote my own little libraries and compared the results to that of sci-kit-learn. I can use books, blogs and wikis to piece and understanding of the mathematics together. So certainly yes in Python. I suspect i could put something together in C to although i wouldn't expect to make best use of the language features and might make mistakes.
My main concern -in general- is that gluing libraries together is not the same as programming, a term BTW that I don't often hear.

And the blind spot (what they don't learn in CS school) is writing code with flexibility and maintainability in mind
 
Last edited:
My main concern -in general- is that gluing libraries together is not the same as programming, a term BTW that I don't often hear.

And the blind spot (what they don't learn in CS school) is writing code with flexibility and maintainability in mind
I also make an effort:
- to adhere to the DRY principle.
- document methods, functions, classes, modules and packages (in Python) using docstrings and markdown files for broader sweeps.
- to adhere to SOLID principles (but i can still improve).

With the exception of algorithms, ultimately I do write quite a bit of glue. Business prefers to get things done ASAP, in principle if the world has solved INSERT into a Microsoft SQL database and it solves a persons problem, then it does not make sense to reinvent that wheel. This said, I do consider myself a programmer, perhaps with less lower level experience that some.

I'd really like to work on a quant project with a real quant, I would in principle do all the work and accept steering. I think if I am to become a quant before too late it will be from a somewhat impressive portfolio. I have made first steps..
 
Top