I'm a first year quant researcher who was hired right out of undergrad by a hedge fund. Obviously not as hardcore or smart as the PhDs I work with, but ask me anything.
>What are your hours?
50-60 hours a week. Closer to the 60 end mainly because I'm a first year and don't want to look bad. Some people definitely work under 50 hours.
>How much of your time is spent testing versus creating a new algo?
Really hard to say since testing is inherently part of the research process. If you're working on a strategy and it has a terrible IR on one asset class, it's totally possible that with a few minor tweaks you can get a pretty good IR on another asset class. That being said, a lot of the idea generation comes from the PhDs, while the junior researchers (undergrads and MFEs) are more responsible for implementation and testing. I still have a lot of flexibility in suggesting tweaks and modifications though.
>What was your undergrad in?
Doubled in computer science and finance from a top 10 school.
>How tight is the leash they have you on?
Less tight than you would expect.
Apologies for the changing ID, I'm on mobile.
No idea. I don't do anything related to HFT or work on any strategies where that matters.
Only the very top state schools like Georgia Tech, UIUC, and Berkeley. Quite a few people who did their undergrad at mediocre schools, but did their grad at top schools.
Applied through OCR on a whim and turned out that I was a good fit based on my background and previous research I did with my professors. I was a pretty nerdy kid who didn't really know how to network, so I got pretty lucky.
What other things besides quantitative finance were you considering doing after undergrad? Or was that your primary goal? I'm in a very similar boat in terms of backround and have no idea what to do with my life.
Machine learning applied to commodities trading.
Pretty typical trading floor setup in rows. Matlab, R, and Python are all commonly used. AWS and in-house clusters available for Hadoop/Spark purposes.
Silicon Valley since I had offers from a few large tech companies with comparable pay. I did an IB internship before, but decided banking wasn't for me. Realized working as a software engineer would be a waste of my background in finance, so I picked this job. More salary growth opportunities as well. Not to mention that a switch from quant finance to tech, from what I've seen, is easier than the other way around. I'd definitely recommend this path if it's available to you.
What courses in CS were totally unnecessary to what you do?
I just plan on taking machine learning and its prerequisite courses.
What courses did you not take that you wish you had taken? The usefulness doesn't have to be tied with your job description, consider the work environment, for instance if all you co-workers were Chinese you would have wished to take Chinese in college.
What internships did you do?
What advice would you give to someone wanting to work in a hedge fund after college?
Is it possible to pursue a graduate degree and work for a fund at the same time?
Are your days long, or is it 50 to 60 hours because you also work weekends?
What are your days like?
All the required classes I took were definitely helpful and I chose my elective classes with a focus on AI/ML so they're all tangentially related. The low-level and EE classes were pretty useless.
Should've taken more math. At the bare minimum, you'll need a strong background in calculus, discrete math, statistics, probability theory, and linear algebra. Classes in convex optimization and the such are also helpful. Sell-side quants do more pricing so differential equations and stochastic calculus are more helpful there. Take classes on financial derivatives (most undergrad finance degrees are a joke and don't teach this stuff in enough depth).
>What internships did you do?
2nd year - IB at BB, 3rd year - S&T at BB
>What advice would you give to someone wanting to work in a hedge fund after college?
Do internships every summer and network as hard as you can. I got very lucky without networking. Buy-side positions available to undergrads are very limited, so don't be discouraged if you don't land one; you can always do 2 years on the sell-side first.
>Is it possible to pursue a graduate degree and work for a fund at the same time?
Never heard of this, but if you find a night-time/online program I don't see why not.
>Are your days long, or is it 50 to 60 hours because you also work weekends?
Honestly 10-12 hours is pretty chill. I don't work weekends.
>What are your days like?
Get in around 7:00am, check on results from whatever has been running overnight, read some news/market commentary, grab a quick breakfast. Code and read academic papers until noon, when I eat lunch, and then code again. Fix any issues with production models during trading hours. Occasional meetings with manager to discuss progress, sell side people to discuss market trends, and other researchers to see what they're working on. Leave at around 6pm.
I would like to work at a hedge fund. Finishing my undergrad this year. Applying to grad school for next year. Ultimately would like to be in the biotech business. Am thinking about doing a MBA healthcare and MSF double. Is this a good combo? Will this combination provide a pathway to a hedge fund or health asset management?
I already graduated with a worthless degree so I know I screwed up. At least you had your shit together. I wanted to do sell-side but without the proper majors degree no one will listen to you.
I'm not experienced enough to give you advice confidently, but I'd say an MBA is preferable to MSF, so you don't even need to do both. Make sure your MBA program is a top 10 or at least top 25 program for a good chance at working in HF/AM. I don't know many good programs that take kids right out of undergrad, so you might want to get a few years of work experience first.
Finance is cruel and superficial, I'm sorry. I hate the rules of the game, but I know they won't change so I just played along.
What's your degree in? I've seen many people get IB jobs with degrees in the social sciences. You just have to network really hard with alumni from your school. Don't just look at bulge bracket firms, also consider boutiques and smaller firms. Otherwise, your best bet is probably to just work a corporate job related to your major and then go for an MBA.
I use math and computer science to develop trading strategies that beat the market. Data can range from financial statements to satellite imagery to online posts.
Most of the things you learn on the job, but the two books I've found most related to what I do lately are Elements of Statistical Learning and Options, Futures, and Other Derivatives. Firms are more concerned about whether you have a good quantitative background and whether you're interested in finance. They'll teach you what you actually need to know.
My major was in international relations, so its a social science. I look at smaller firms, since its my best bet to learn the trade. But I don't know how to write a resume related to finance.
With respect to your experience in the industry, how viable have ANNs for quantitative analysis and forecasting within different time frames?
In the simple (hobby) simulation I set up, they seemed to consistently outperform the market when there was a stable upward trend, underperform when there was a stable downward trend, and freak the hell out when there was excessive volatility. At the time I got distracted by other life obligations, I was investigating ways to anticipate market conditions during a given day to improve risk management with respect to the performance patterns I observed.
Thoughts? Any similar observations/insight?
What Industries does your firm focus on like general info Health, Retail, Tech etc. If you do focus on any, does your firm hire analysts that have an expertise in those industries.
Also, I'm hoping to break into sell side finance or MBB consulting. Do you have any tips or inside information that you can a Finance/ Info Systems student. My financial knowledge is good, Maths is ok & Interpersonal skills are great.
Read http://www.mergersandinquisitions.com/ and http://wallstreetoasis.com/ for information on breaking into the sell-side. It'll be difficult, but international relations honestly isn't too bad of a major for finance. I've seen many people pull it off.
Neural networks and deep learning are amazing for many things, but from my experience in industry they're not great for systematic investing. The consensus, at least at my firm, is that they're too much of a black box and the weights/gradients are not easily interpretable when compared to more traditional ML techniques like SVMs and decision trees. The theory isn't as well understood and there just isn't much economic/finance intuition behind the results.
Extremely important for understanding theoretical computer science, which is where I feel quant investing is moving towards. You'd be surprised how useful combinatorics and graph theory can be when developing your own algorithms.
I kind of regret dedicating all my time in school to studying and polishing my resume. Really wish I spent more time having fun and being social, but I'm sure many other engineers share the same sentiment. Didn't want to be an actuary because it seemed repetitive and I've heard salary isn't very good until you pass the exams.
We don't have a particular industry focus. I actually work mostly with fixed income products rather than equities. Read the two sites I listed above for advice.
How low should the Eout for a regression perceptron be for practical use?
Would it be wise too feed your strategy with regression & logistic regression input to strengthen performance?
Assuming u want to trade on a 1-5day timeframe, how often should you retrain your networks with the new data?
NN's or SVM's for timeseries forecasting?
>I use math and computer science to develop trading strategies that beat the market. Data can range from financial statements to satellite imagery to online posts.
First year Auditor here, I prepare those financial statements documents that you use. I've always been curious about investing and this might be the year.
Question - what exactly do you look at in regard to the financial? Equation? Please, I'm very interested.