I am almost finished with my master's in compsci, focus on machine learning. I tend to want to work on general artificial intelligence, but then I feel like just inventing this shit for you chads who will use it to get even more rich while I get status points in the "academic community" which I already hate (conservative unimaginative petty cucklords mostly).
So the question is: How would you go about maximising your chances of obtaining a few million in my situation? Bonus points if I can continue working on AGI or don't have to suck dick all day.
Ideas I came up with: - tech startup (unlikely to work do to low apriori chance combined with me neither a coding nerd nor a complete expert) - whoring myself out (low expected payout due to reason above) - try to improve my skill and become master AI wizard; maybe by working in uni research while continue studying at home (chance of wasting a lot of time, aint getting younger. Also danger to motivation if payout to plan is less secure due to more possible complications) - as above but more focus on practical trading algorithms (small chance of success only and less growth in AGI skill, just bland implementation/modification of existing algorithms)
why not try to get hired by some buyside firm currently using AI... I mean aside from google deep mind, Facebook and Microsoft research the biggest employers of machine learning researchers outside academia are probably hedge funds and prop firms (I don't mean generic 'data scientists' with some quantitative post grad degree + some knowledge of stats/ML but people with specific ML post grad experience working as actual researchers)
game to play then is generally stick around somewhere for 3 years, wangle your way out of the non-compete some how then negotiate your way into another firm on the basis that you're now bringing models/alpha generators with you... i.e. you own the pnl now and get a fixed % of profits rather that an discretionary bonus. This can also lead to situations where you're almost self employed - you might want to protect your own contractually IP/maintain ownership of it - this can even lead to situations where some traders/PMs maintain ownership of the servers used for the strategy themselves in their own LLC or ltd company in order to protect against their new employer stealing the IP and sacking them
Jaxon Green
millennium and world quant are the sorts of places where you can go for the quasi-self employed route... but the key is to start somewhere else first and learn from an experienced team
there are only really a few dozen strategies being used to generate alpha out there - everyone is basically hiring researchers to make tweaks/adjustments to the same set of ideas that is why you'll see clear correlation in the returns of certain funds and CTAs - especially when certain people have been behind founding them
Jack Walker
Thing is (this might be impostor syndrome, but the flipside is dunning krueger), I feel like having a big scar in my face due to not having done any practica (is this the word? sorry, am German. what I mean is working for low wage during studies to gather "work experience) outside of university at least, taken one year longer than generally expected and having a bad gpa for the bachelor's (master's seems to be good though).
Also I don't understand most of your words in paragraph 2, sorry. But I gather and agree that I have achieved the first base when gotten hired by some good company.
Isaiah Brooks
Join a bank or hedge fund as a quant, you can have 500-700k saved up within 5 years.
Hunter Watson
it might not be feasible then... the above is sort of reliant on your initially landing a junior researcher role which is hard enough to do without a PhD but still possible with a good degree from the right place/good grades etc... with that sort of role you're researching trading strategies within an experienced team (well some places are more collaborative than others) and have access to all the data and infrastructure you'll need
trying to look at trading strategies as a complete outsider with no professional experience or others to learn from is a massive uphill struggle, there simply isn't much published material to help you - almost anything published about the subject itself is utter garbage. Sure you can learn about financial instruments, learn about market micro-structure etc.. but stuff dedicated to actual trading is generally crap
you're probably better off going into the tech industry and aiming for data science roles
maybe consider becoming an actuary - old established profession, pays consistently well, fewer hours than finance and someone probably could add some serious value there with new approaches from an ML background once established
Nicholas Gomez
Ok thank you, looking into it. What I feel a little afraid of that I was socialized with nerds and sjws, so I feel that I might be classified as outgroup when going there. Also my image is mostly from wolf of wallstreet, so I might be an idiot here.
Jaxon Kelly
hey OP, i am somehow in the same boat as you: graduating in 2 months with a M.Sc. in Statistics with a machine learning focus and still looking what to do afterwards. A PhD is not that interesting to me as I'd rather do something more personally fulfulling such as starting my own startup or something along those lines. How about staying in touch? (German also)
Logan Hughes
Well thing what interests me so much about AGI that you don't need expensive experts anymore with it ideally, the algorithm can gather it on it's own.
Pic related; few years ago every shortsighted guy at university told me "machine learning is about manual feature engineering". Good that I did not take them seriously, but it was a close one.
Adrian King
wolf of wall street isn't an accurate depiction of characters you'll find in a bank - it wasn't based on a bank anyway just a shitty boiler room
if you want a realistic portrail of the characters and personalities you'll find then watch margin call. Jeremy Irons plays the CEO role perfectly, Paul Bettany's character is pretty realistic for a sales trader and Zachary Quinto is a very realistic quant - credit to them for not going for some silly Hollywood nerd stereotype.
