Hacker News Reader

Comment by turdprincess | Hacker News Reader

You get excited for one reason and one reason only - the money. FAANG jobs pay between double and triple of normal software gigs. Some folks want to optimize for interesting work, or groundbreaking products or autonomy. Thats cool - you might not find that at a FAANG. Some folks optimize for the biggest paycheck. For that its basically two options - FAANG (and adjacent) or High Frequency Trading. To each their own!

View on HN

I used to believe that, but searching for a job now, I see many non-FAANG companies with compensation in-line, or better, with the exception of those who aren't public. OpenAI for example, has base salaries in the $400k range for ML jobs at $250k range elsewhere at like an IC5 level, because of non-public equity.

Yes, I think FAANG grouping is too narrow. The true group is "big tech" - with the regular FAANGS, as well as Uber, Airbnb, Stripe, Pinterest, etc...

While each company has some unique qualities, my guess is that being an engineer in any of these places would be pretty similar and include:

* A lot of engineers and teams

* Lots of organizational / promo based manuevering and social engineering

* Problems at scale - some technology, but mainly "number of cooks in the kitchen"

* Work which might not be very interesting

* high compensation

My guess is that the OP's concerns would hold true at most of these big tech places that pay top dollar.

There are other reasons too like scale. Obviously not everyone can work on those types of projects but few companies can claim their engineers' code will affect a billion users for example. At that scale, the problems that need to be solved can be interesting even if they're trivial for a smaller set of users.

How does one get into HFT?

I think your options are:

* get a CS degree from a top tier program, preferably a PHD

* Pivot in from a normal software job. This is probably only possible if you have a bunch of C++ experience, and some top (FAANG) names on your resume.

Learn low latency network programming in either C++ or something safe like Rust. If you make something people notice HFT companies will beat a path to your door.

Learn Rust, C++, Haskell or OCaml. Or get a PhD in mathematics or physics. Other languages or domains might work, too.

Rust mentioned twice here, maybe there’s something too it

I don’t think it’s possible for me to go down a PHD path.

HFT seems like a cool gig if you can get it

I've been approached several times for HFT or Quant jobs (yes, I know that they're different but I tend to sort them in the same bucket in my mind) because I code in Rust, C++ and OCaml.

I didn't go into the domain because I'm not a big fan of money games. YMMV.

Reply on news.ycombinator.com