Education required for algorithmic trading
Hello all, I’m a freshman in college currently pursuing a BBA in a top public program and planning to dual degree with computer science on a data-analytics heavy track. As I understand, the expectations are quite different for different “quant” roles, and I’m a little confused on what is actually required to land what I’ve seen called a “trader” role in a hf that develops code to execute strategies found by usually phd researchers. I’ve been told by some that my undergrad program is enough, others have said an MFE or other masters degree is expected, and others have said I should even have a phd. I don’t want to waste time and money on education that I don’t need, so what do I really need to do to get an interview for a role in this industry?
Undergrad is enough for a coding role. Develop codes to execute strategies is a quant developer role, not trader role.
Traders still need to know how to code in prop shops. I think the large majority of traders come from computer science backgrounds
Not really. Discretionary traders don’t write codes. Quant traders write Python and R codes to research strategies, but those codes are very different from the codes used in execution. Codes and codes are different
you're literally dead wrong. I don't even know where you get the arrogance to reply when you clearly do not have an understanding of quant or what they do. The developers/researchers research the strategies, and the traders write scripts to execute them. It will vary from shop to shop but the idea that quant traders and using R to run statistical analysis on their own is complete idiocy, that's not their job.
What you said is just plain wrong. Researchers research strategies, developers do a lot of things: build trading systems for execution, build research platform, build data pipelines, implement ML algo provided by researchers ( use low-level language like C++ and Java to achieve low latency ) . Quant traders tweak parameters of the algorithm in reaction to changes in market conditions ( using scripting language like R and Python ) so quant traders are also doing research but they don’t build the entire algorithm they’re more on the tweaking side. Developers build execution softwares and toolings for quant traders and researchers. If you’re talking about execution trader, those people are already 95% automated by softwares. By “trader” I’m clearly talking about quant traders, not execution traders. Also developers do not research strategies, developers are coders and engineers who work closely with researchers. It’s clear that you know zero about this industry, I don’t know where you get the courage to bullshit here.
And if you’d even bother opening the career page of Jane Street and look at their job description for quant trading, you would have known you’re wrong.
Note: in some funds, all of these roles are mingled together and the boundary of roles get blurred. For example, developer doing research, trader writing code, researcher doing execution. People don’t like been staggered to one role so they switch roles all the time, but in most top-tier quant funds, these roles have very clear boundary because the fund do not like people stepping on each other’s toes
At least some HFT firms prefer to have their algorithmic traders be able to write C++ or whatever language the firm uses in production so they can make some relatively minor tweaks to production code if they do some feature engineering or make some simple changes to execution logic although this would be often be code reviewed by a developer. Most of the trading algorithm like market data parsing, basic execution logic, and risk checks will typically be handled by dedicated developers. There are also almost inevitable tradeoffs between feature/model complexity and being fast and it helps to understand both the trading side and the dev side when navigating those tradeoffs even if you aren't implementing the production code yourself. For less latency sensitive trading strategies the importance of this is much reduced and if you are really talented at understanding markets or doing quant research you can still do well in HFT without touching the production codebase but at many firms your life will be easier if you can do some code changes yourself.
So traders tweak the algo prior to execution? That's basically what I said, they're not exactly researching strategies. fwiw I misunderstood you when you said that traders are researching the strategies, I guess we both know that's not really true.
I will be working in quant trading this summer and have shadowed some friends, almost all of them come from extensive CS backgrounds and have to have a real in depth knowledge of CS in order to perform on the job. jane street is fairly particular in the way the traders still make a lot of discretionary decisions.
Developing code to execute strategies is something I'd more likely see in a quant dev than in a quant trader, but different firms work differently. Undergrad is sufficient for dev work. I really only see MFE/other grad school being worth it if you wash out of recruitment your senior year and want more time to apply for entry level stuff.
I joined my fund straight out of my undergrad degree as a quant trader. As a trader my job is to apply mathematical models to the market depending on what I see fit to be the current market dynamics, a bsc / Msc is more than enough for this if your coming from a strong maths background. Sadly for you, cs is just not considered mathematically heavy or rigorous, normally if traders have a cs degree it will be a joint honours alongside maths (which I think is the perfect degree for the job). Where the cs people run the show is in HFT where the actual strategies being deployed are relatively simple but speed is imperative so the understanding of writing super efficient code is required. The problem is that algo dev at HRT may be the hardest job in the world to get out of university and going to work for no name company with no history is just a bad form of gambling. The job title in this industry means completely different things depending what company you work for, I would advise if you arnt exceptional at maths and coming form a cs background to try and join a company where your role is similar to the following... Someone who's only job is to generate alpha signals who works in python or some other high level language then hands you the strategy template for you to design and try and optimise in low latency language.
My school does also offer a quite good bs in data science program that I could switch to if need be at this point, because the first year or two in the two programs is nearly identical. I would still take all the core compsci courses as well as some simple machine learning, just with datasci electives instead. Do you think that switching to that from cs would better equip me, or would it just hinder my swe skills? I like the sound of the role you described but it also sounds like it’s very niche and might be a hard position to come by.
Not particularly related to the OP's question because I agree CS is more useful in HFT than mid or low frequency trading but I don't think the strategies deployed by quantitative HFT firms like HRT are actually simple even if the amount of computation done in the fast path is relatively little. Compared to mid or low frequency there is certainly more emphasis on the engineering aspects to keep strategy latency low even when the model is very complicated but it's also arguably easier to avoid overfitting in HFT so to the extent strategy latency is not too impacted there can be significant advantages to rather large and complicated HFT models. Whatever simple but super latency sensitive models that are still competitive in US markets are also very likely entirely in hardware and typically developed by specialized hardware engineers instead of software engineers.
Consequatur eius quaerat sint voluptas. Qui ut sed ipsum non debitis quos iste. Dolores enim dolorem ipsum nostrum et. Delectus eos ea ea veritatis.
Aut consequuntur et dicta nulla sit aperiam. Est sed in quo ad labore tenetur hic qui. Et porro officia odio.
Voluptatem non perferendis soluta asperiores modi non. Et nam aut et. Omnis modi qui aliquid et quia.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...