A few months ago, I told my boss I'm interested in AI. She indicated last week that we have a small (<4 month) project on the horizon which involves feeding an AI model all of our troubleshooting guides and data queries for oncall work, which will then help categorize oncall tickets as they arrive along with doing preliminary queries for us. I honestly haven't studied AI in a few years, but I understand that this will involve LLMs and NLP. I'm currently reading picrel as a primer to get back into AI/ML in general. What would you guys suggest I read/watch/whatever to ramp up on LLMs, NLP, and whatever else you think I'll need for this project? I have a few months to prep.
Ask an AI
>she
aha good luck bro
sounds like what theyre looking to do is basically using openai apis to make oncall work easier. why are you studying machine learning for this? don't really seem related.
Are OpenAI's APIs really that simple? Shouldn't I be at least somewhat familiar with how these things are working under the hood before I spend a quarter of a year on a project with their tools?
It's Meta. None of you brainlets could get hired here.
OpenAI API calls are literally just giving it a text prompt and getting text back. The ""hard"" part comes when you want to do RAG (being able to query other sources with LLMs), which you definitely will need in your case.
Learn the basics of feature learning and autoencoders to do vectorization on your support guides, which is the part that comes closest to ML. The rest though? It's literally prooooompting, but do get used to MLflow or any experiment tracking tool and some way of testing prompts systematically that's not just "uh just reword that part bro".
>calls others brainlets
>can't even figure out where to start learning in the online era
have a nice day
Sorry, didn't mean to interrupt the tech discussion board with a discussion of tech. Don't let me keep you from gushing from every orifice about the 'eck or whatever it is you think we should be talking about right now.
there's absolutely nothing wrong with discussing tech here, but don't expect help from the same people you are insulting, moron
>It's Meta
as a fellow moron, can you give tips on how you got hired there?
Getting the interview is the hardest part, honestly. You need a good resume, meaning some combination of good school, notable personal projects, internship experience, relevant skills, etc. As far as acing the interview, you really just need to grind LC patterns until you can solve mediums easily. Hards are a crap shoot, and anybody who's telling you they can solve most hards without having seen them before is either lying or is some kinda savant.
Depending on level, you might need to know system design too, but you won't be expected to know anything you can't learn from that DDIA book. To be fair, you don't have to be smart to get hired here, but if you're the type to gloss over details when you're reading documentation or when learning a new algorithm, you're gonna have hard time studying what you need to study to do well in your onsite. I'd recommend doing the LC 75 then doing extra questions for the topics that you don't really understand.
Also, you there's no point in getting hired here if you're just gonna get pip'd in a few months because you only know how to LC. Personal projects should cover this, but you gotta be capable of pushing quality code and holding your own when oncall, or you're just gonna end up wasting a bunch of time interviewing for jobs you can't actually do.
FAANG devs aren't geniuses, but anybody who thinks the average dev here is lazy or stupid is coping. You really, really can't afford to be lazy here. You won't last long.
didn't expect a good reply, thanks
and it doesn't even sound impossible tbh
I didn't consider faang realistic until now but maybe if I stop being moronic and do an internship,,,
>have 11yoe
>be yuropoor
>used to participate in coding challenges in my undergrad
>been through six million interviews to get to the place i am at right now
>working 9am-11pm mon-sat
>still haven't sent my CV to gayMAN
>this is all it literally takes
what the frick am i doing with my life
>9am-11pm
the frick are you doing with your life indeed, I seriously hope they're covering you with gold because those hours are nightmare fuel
well first you must start with calculus
once your learn derivatives, you can move on to integrals
and from there move to linear algebra
once you learn matrices and stuff, you'll want to learn probability
after that you should learn a good amount of python
also a good idea to study oop and other programming concepts
after that you should start professor N(i)g(gga)'s coursero course
after th
just make the thing
>his boss is a woman
EXCUSE ME I WORK AT **META**?
LIKE, HELLO???
memeing is supposed to be funny anon, not just poorly disguised butthurt, nta even
MEMING IS SUPPOSES TO BE FUNNY, OK???
I tried to buy some ml books, but they always was out of date when they arrived. So gave up.