Becoming an algotrader
Hi everyone,
A bit of a beginner's question.
I am a current engineering major at a top 10 worldwide school with a very strong interest in finance. I have secured a spring week in trading at UBS, and hope to convert it to a summer, and a summer internship as a finance analyst at big firm (unrelated but good for experience) - I am aware this is not related to trading or IB in general, but the opportunity arose and banks do not take people who are not in the penultimate year of their course.
Anyway, recently I have decided to try and pursue a career in algotrading. But my knowledge is extremely limited. I only learn MATLAB in my course, and, whilst I take a couple of finance minors, none of them are related to computational finance or anything similar.
I was therefore wondering what you would recommend a college engineering major to do in order to get on track to becoming an algotrader. I am currently self-learning Python at a fast pace (started a few weeks ago and I'm quite comfortable with the language). I am very well-prepared in mathematics and statistics and ready for long hours of work to prepare, but I have no idea where to start.
Of course, I would do a postgrad - potentially in computational finance or financial engineering. How necessary is a PhD? How can I break into the world of algo-trading? Any online or summer courses you'd recommend?
Thanks!
Isaac
What are you using to learn Python? Any recommendations?
Majority of algorithmic trading houses will require solid undergrad grades from a top uni, + a statistics/maths intensive masters. The jury is out on PHD's. A PHD will obviously be a positive (no question about that) but that's not to say you can't get in without one.
It's an interesting field - I don't work in it, but am interested in the area via my work in physical commodities and considering a masters myself.
This website introduces you to basic Python, fairly straightforward: https://learnpythonthehardway.org/book/
I am definitely doing the masters. The question lies on the PhD, and landing a job in algotrading without any prior experience in a firm, as no one is going to take a college undergrad for that.
Learn Python the Hard Way is a great resource, I personally used it a few years back. When you have the basics covered be sure to check out Sentdex (sorry, can't link to youtube since my account is new) for more finance and statistics specific tutorials.
Essentially I am going to recommend you become more of a generalist than the areas of study you talked about in your post. What interests you now may not be interesting to you in 5 years, so I recommend getting the skills you need for quant finance, but doing it without pigeon holing yourself into only finance roles.
What sort of engineering are you studying now? I ask because some types naturally lend themselves to finance more easily? If it's civil/environmental engineering you'll likely find it more difficult than EE. It's great if you have time to tack on a minor while you're still in undergrad, but you'll be better served making it computer science or mathematics (probably in that order of desirability) instead of finance. You can develop sufficient finance knowledge on the side. As for grad school, assuming you're doing well in undergrad and are taking sufficiently technical courses in your engineering program, you should be competitive for graduate programs in computational/applied mathematics, operations research, or even computationally-aligned engineering groups (eg if you're ME/AE and are good at CFD or FEA). Any of these sorts of programs will give you a good skill base that you can flesh out with some additional finance/math knowledge when it comes time to find a job.
Finally, if you're considering a PhD program, but not planning on pursuing academia, I will go against the grain and recommend weighting your decision on where to go slightly less on your individual adviser and slightly more towards prestige/brand recognition of the school. It's very likely in finance nobody will have heard of your adviser either way, but for better or worse most recruiting into finance roles comes from the same 20-30 schools.
Thank you very much for your reply. As you mention, I do not want to pigeon hole myself into a very specific role. I study chemical engineering; we have quite a lot of computation, but it is all in chemical engineering-specific languages. My degree is technical and I am maintaining a 4.00 GPA average; so hopefully it leaves me in a good place for my postgrad. I am very prestige-focused, so I'd only aim for top schools rather than top programmes at not-known schools, as they normally have more resources and open more doors when it comes to networking and meeting people. How do you think I can blend all of my knowledge (programming, finance knowledge and mathematics) together? Would this be in my postgrad? Or through personal research? Thanks!
That's great that you have a good GPA, it's definitely important when it comes time to apply to grad school. I think from a ChemE background you have some decent opportunities to work on interesting projects that could help you land a role in finance. I am not super familiar with current research in the field, but I would guess that there are substantial computational modelling challenges in biological/physical chemistry. Ideally I would research faculty at the schools you are considering in computational/applied math, applied physics, Chem E, and perhaps materials science to see if they have research projects that would allow you develop a more rigorous math/modelling/data skillset while complementing your chemE work. If you do a minor in math/CS you may also have an opportunity to do grad work in those fields while leaving behind the CE stuff altogether. All of this is predicated on the assumption you are going for a PhD though. If you only want a masters I would recommend an MFE or similar program, although I think you're going to have more trouble breaking in to algo trading from that path.
If you learn about markets and finance on the side, it will be sufficient. I have interviewed plenty of people with PhDs in non-finance fields. The ones that do the best are those that can carry on a conversation about the macro environment and especially volatility, even if they haven't ever traded or worked in finance. If you can do some side projects that's even better, but I don't think its required. You just need to be genuinely interested. If you are you will enjoy researching markets in your spare time and the knowledge you develop will come though.
Can someone explain me the logic behind majoring in Chemical Engineering and pursuing algo trading that is purely computer science driven? It like I want to be a writer, but I'm going to study cooking to become a better writer.
Generally the more you derive edge from latency in the market, the more CS driven the job is and the less chance a general "problem solver" has to make contributions on an algo team. That said there are plenty of markets and firms where the primary edge comes from other sources and they can use smart people who are good at thinking about and solving hard problems in general. You'll still need some programming skills, but the job becomes more about the ideas you generate and less about the perfect technical execution of those ideas. My feeling is that those are the sorts of jobs OP should try to shoot for.
I disagree with most of what has been said. If you want to be an algo trader you better have a lot better and in depth coding experience than self-teaching yourself Python. You should be able to write extremely complicated algorithms based on financial theory and mathematical models that can handle the complexities of the real life markets. This is usually done in C++, Python, and others. You should be comfortable with many languages. Ideally, you would want to have a bachelor's in computer science with an minor in math, statistics, economics, quantitative finance, or finance. Seeing how that is not ideal given your current situation in engineering you could potentially do a masters in computer science or find a way to apply your skills you have already learned in coding and put them to the test working a full-time job where you can design some algorithms related to finance, but where you are not necessarily making them for the trading floor. I.e. working in operations or the technology division of a big bank. Remember if you get the job you are being hired to trade millions of dollars at a time. They need to be able to trust you and honestly I wouldn't trust some random kid who taught himself how to code so you need a way to prove to them your up to the job by finding a job where you can build the skills you need. A masters in MFE won't get you what you need. You need either a masters in math, stats, or computer science paired with significant math exposure. If you want to be a quant you need a PhD. I would try to get a job in the middle office of an investment bank and work there a few years till you can learn the basics of trading how it all works. Go do a masters and then come back after finishing up grad school. If you build the right connections in your post-bach job you should be able to do it.
+1 SB
Well said.
I don't really think this is true at least in HFT. Most HFT algorithms aren't extremely complicated because of the need to be very fast. Almost all firms use C++ for this. The research process/parameter fitting is different as there is more time for computations so Python is more popular. I also think most algo-based prop firms use very little financial theory.
HRT, KCG, Jump, DE Shaw, Citadel, and Two Sigma among others regularly hire undergrads for quant roles who know little to nothing about finance/trading. These firms are all quite selective and it is possible that other firms (banks?) only hire PhDs for quant roles.
apply to summer internships in sales & trading programs at every BB (there are 20). then apply to 1st year analyst programs at every BB (yes, all 20).
take the best offer you get...and learn on the job. you can't learn the good stuff in school...because the teachers don't know it themselves. only traders on a trading desk can teach you that
Getting into algorithmic trading (Originally Posted: 09/01/2010)
Hello forum users,
I was hoping for a bit of brutally honest advice. i'm currently a big 4 trainee accountant in the wrong job (UK based). I'm highly numerate - Maths Olympiad level, and autistic (really helps with maths) . After realising I was in the wrong job I looked up people I respected and met up with them to describe their job to me. One guy worked for MTH and gave me a lowdown of how and what he did for a living and it sounded like an awesome fit for me. It wasn't a "i could do that job moment" it was a "that job is perfect for me" moment.
I'm a physics grad from a top uni and a few interesting points on my CV, ranging from a qualified ski instructor to starting my own business that has a way of beating the gambling industry (point i'm trying to make there is that I don't cold call vulnerable people to make money, and can identify ways of exploiting weak points in systems).
I was recommended and thoroughly enjoyed a lot of books, black swan, Liars Poker, When Genius Failed, fooled by randomness. I'm contemplating getting Options future and derivatives by Hull, for 170 pages i'm not sure what harm it can do. (any other book recommendations?)
I was going to dry run my CV on a couple of institutions because I really don't want to screw this up. However i'm concerned about a couple of things that I'd like some direct and honest opinions on.
I was during my years, less so recently, a very high level online gamer (strategy only). I attribute this to being able to look at a system, understand its weaknesses and exploit them. I feel that this could be relevant in an interview but I'm concerned it could come across oddly. Was also the leader of a high end raiding guild on WoW, is that too far? (i've long since quit). I want to show strong performance under pressure, and also a bit nerdy/geeky.
Also I don't feel comfortable with my answer for "why do you want to work for us as an alg. trader. It's not that I'm making up some random what you want to hear answer, i just struggle to express whats in my head. Are there any definite does and don'ts for this question?
I'm also aware that I'm not so clued up on trading, beyond the dummies basics. I've been told that they teach people who are new so industry experience isnt a problem. Should I be focussing on getting more knowledge on this or is my time best spend elsewhere?
Finally, aside from MTH, are there any other London HFT firms that you would recommend as suitable for me?
Any feedback is gratefully received.
Regards T
list your programing skills c++ java pearl python ruby c#
I've not really had a chance to use my programming skills after university. I'm pretty strong with excel since I use it a lot, but my C++ days are a while back. Is there anything I can do to get back to proven skills without going on some course lead by someone that speaks Klingon. (the firm I want to get into just requires one of excel c++ etc.).
Any more thoughts about the rest of the CV?
take some CS classes in c++ and java in CC
Trading Algorithms (Originally Posted: 04/08/2009)
Could anyone provide just a simple explanation of algorithmic trading? I understand that they are basically programs telling you when to buy and sell but I would like to get a little deeper understanding than that.
http://lmgtfy.com/?q=algorithmic+trading
hahaha. that was awesome.
lol i laughed....Statistical arbitrage, Market neutral, and different types of relative arb strats look them up.
lol hahaha good one
classic. well played.
Only two things are infinite, the universe and human stupidity, and I'm not sure about the former.
Oh man that's good. Just use wikipedia dude.
Starting off as algo trader - advice? (Originally Posted: 12/25/2010)
I will be starting off as an algorithmic trader in Chicago this summer, right after finishing undergrad in Boston. I was wondering if you guys have some advice for me to make it on the job, as well as when it comes to lifestyle and anything in general. I've heard and read so much about making it on the interviews etc, but pretty much nothing on being successful once you're there. Any recommendations about resources on this would be appreciated as well.
Very important to clearly outline objectives before you start programming. Are you working on a team of programmers or are you solo?
Well, I will be part of a team, but I will design and implement my own models and algorithms on my own. There is a process of review and feedback, of course.
there's a quant forum over at
http://www.wilmott.com/index.cfm?NoCookies=Yes&forumid=1
that might be more helpful than WSO for you
creating a trading algorithm/program (Originally Posted: 07/07/2016)
I'm looking to learn more about programming/creating algorithms to run trade ideas I have. how do you go about creating a program? for example, where do you get data feeds from to get information for your algorithm? and how is a trade executed based upon this (is there a platform you set your program up with?)? I apologize for my lack of knowledge. I am new to this and any help is much appreciated. thanks!
Well you find a broker that provides API (kind of Interactive Brokers) and the you connect to that API and do your trading. Here is an example for Python https://www.quantstart.com/articles/using-python-ibpy-and-the-interacti…
while(true){ if(market.isDip()){ buy(); } }
Quantstart will have the majority of the info you need to get started
There are many online resources to make use of:
Quantopian - get started with built-in trading strategies, start trading immediately QuantInsti - download free models and trading strategies in Excel/R/Python QuantStrat - learn about algorithmic trading strategies
Quantocracy - hub of all latest Quantitative blogs and content
Self-teaching Algorithmic Trading? (Originally Posted: 10/10/2017)
How feasible is it to self-teach algorithmic trading if I only have knowledge of up to Calculus 2. I'm currently a sophomore in university and haven't declared my major/minor. I might do a computer science/mathematics minor to help me get up to speed, but other than that I would probably be reading books and coding.
Thoughts?
If you're interested, major in Computer Science. Keep your options open. The positions available for algorithmic trading are slim and among the most competitive on earth.
Consider Quantopian in your free time once you have mastered algorithm analysis and prepared for traditional engineering interviews.
Dicta nostrum in rem quo. Molestiae dolores veritatis voluptatem. Repudiandae non sunt sunt ex similique culpa.
In impedit et repudiandae rerum beatae saepe. Quod dolor recusandae aut sequi nostrum. Quidem nostrum enim tempora assumenda.
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...
Distinctio magnam unde magnam laudantium. Ut dolore sed voluptatum aspernatur et et sint sit. Aut aperiam delectus excepturi explicabo velit sed adipisci ipsum. Harum cupiditate ratione nesciunt cum tempora.
Voluptas error fugiat voluptates. Beatae soluta maiores dolorem vel. Sit voluptas et sint mollitia. Ut ducimus omnis quae harum optio pariatur quasi.