Hustling into Algo/HF (Quant)

HF

6y

malloc:
What is the current HF "hot" problem? What can I drop in front of a PM and have him salivating to fela- hire me?

My guess is that most things you can come up with as a hobby are not likely to get anyone excited. That's especially true with anything high-freak-related you can come up with. So no, don't get all immersed in C++ and start squeezing those mikes in w2w latency. Not to say that there is no value in knowing that, but that's not a skill that would get you employed these days.

If I had to find a something sexy that would attract attention of a quant PM ("Personified in this case by an 'orrible cunt ...me"), I'd look at applications of alternative data (non-market), especially involving machine learning or fancy stats. In the process you'd acquire a relatively rare skillset (CS + data science). People like that are well bid outside of finance so you will have a competitive advantage even if you change your mind regarding the whole quant thing.

I have a friend who lives in the country, and it's supposed to be an hour from 42nd Street. A lie! The only thing that's an hour from 42nd Street is 43rd Street!

ooh ooh aah aah

O

6y

.

Accredited Investor Definition 2022

HF

6y

Best Response

malloc:
If 1/3: I probably will never be able to go toe-to-toe with dedicated teams, but I can get close. Maybe not nano-seconds close, but optimizing the shit out of an FPGA and making a "no-kernel" kernel will put me in the microseconds range.

It's more along the lines of 2 and the fact that low latency engineering has been commoditized over the last few years (for various reasons). The times when you can get paid well for simply knowing how to build fast systems are more or less gone.

malloc:
If nothing else, it'll be something to point at when asked for proof of my skillset. Because after looking through some of these quant-dev requirements (Jane, Citadel, HC, BlackRock), I know I've got all the dev experience needed, but nothing nicely wrapped up to prove it.

No, you do not, trust me on that. Not because you are not smart enough, but mainly because you don't have the domain experience. But yes, if you build something like that it would help at an interview (even though I am pretty sure any HF developer will tear you apart anyway).

malloc:
I want to focus on HFT/dev work, because it's what I'm good at.
Research/financial engineering on the side. So if dev work doesn't turn out to be challenging/rewarding enough, I can attempt to switch gears and go MFE without being completely blind-sided.

Firstly, you probably don't don't really know what you are good at, especially at this point in life. You tried software development and "it works", so you think you are in love with it. Once you have to maintain a pile of noodle code (written by the likes of myself), you might reconsider.

Secondly, there is no realistic way to do research "on the side", especially if you are busy maintaining the plumbing (that's what most developers are doing 90% of the time). While you will have some minor exposure to finance, you are not going to gain the skills needed to produce alpha.

I have a friend who lives in the country, and it's supposed to be an hour from 42nd Street. A lie! The only thing that's an hour from 42nd Street is 43rd Street!

HF

6y

malloc:
I thought long and hard about your posts, and as much as I wanted to "fight back," I know you're right.

Thanks! As an aside, the first thing you learn in this business it's not about being right. Personally, I'd be happy to be called an idiot 3 times a day if it doubles my PnL.

malloc:
What do I need to know or have, so that I can hold my ground during the interview -- atleast compared to all of the other candidates? Besides the implied interviewing and technical skills, is it just weathering the blows to my ego?

To quote from myself a little while ago:

Mostly Random Dude:
Here is a list of what I looked for when hiring recently: - have at least one programming language that you are very comfortable with - two languages, one prototyping/research and one for power computing are nice - be comfortable with most statistical concepts, it's surprising how many people aren't - basic linear algebra and surrounding data processing ideas (PCA etc) - most importantly, have a brain - that means know what you don't know

malloc:
It's not my first choice, but it's the only one where I've built up enough discipline to keep on grinding when it starts getting into the "troughs of sorrow."

Well, that's the crux of the question - what is your first choice?

I have a friend who lives in the country, and it's supposed to be an hour from 42nd Street. A lie! The only thing that's an hour from 42nd Street is 43rd Street!

ooh ooh aah aah

O

6y

.

HF

6y

malloc:
I appreciate you humoring me, and I've got a crystal-clear picture of what I need to do now. The last thing that's on my mind: how do you not let fear and uncertainty get to you?

This is scary, I literally was just talking to a someone about it. Here are a few thoughts

(a) On the balance, fear is a good thing, fear is there to protect you. So as a risk-taker, you should embrace a healthy balance of fear and greed. In general, good entrepreneurs or traders either have that balance themselves or want to work on a team that has the right greed to fear balance. If you know that you personally have a tendency to be either overly cautious or overly optimistic, you got two choices. You can try to find a partner or an employee who is your exact opposite. Alternatively, (this is my approach), whenever faced with a decision or uncertainty, write down your thoughts on both sides and make sure that you have equal.

(b) Most humans are not really equipped to think about risk in a priori terms, but rather tend to think of risk a posteriori, once the result of the decision is know. Here is an example. At a Super Bowl a few years ago (2015?), Seahawks coach Pete Carroll called for a pass, with like a minute remaining on the clock. The pass got intercepted and the coach got a load of shit for it. Like newspapers were calling it "the worst call in history" and such. However, statistically, passes get intercepted very rarely and probabilistically it was the right decision.

(c) My general approach (to trading and life) is to evaluate the worst case scenario (and come up with some sort of chain of contingencies), the best case scenario and, most importantly, the median outcome. Median outcome is more important than mean outcome for most of your life's bets - that's what Kelly criteria is all about. It's surprising that people have a good enough idea on what to do in cases of good luck, plenty of people think of what to do in case of bad luck but most people do not consider what to do when the luck is average.

(d) Risk management is not about trying to predict the exact nature of disasters. In fact, people who do that (and boast about it forever after) are the worst enemy of proper risk management. Imagine a military commander that says "the enemy is going to attack from the West, let's put all our forces there" - is that good risk management? So it's more about finding gaps in the current setup and coming up with a process should something unexpected happens.

(e) Last but not least. If you are taking risk, fuck-ups are inevitable. It's the decisions you make after that matter.

Anyways, I can talk about this for hours so let's leave it at that.