Becoming a Research Engineer at a Big LLM Lab | Hacker News Reader

Becoming a Research Engineer at a Big LLM Lab

maxmynter.com

55 pointsfelixdoerp21 comments

My last job search 2 years ago took me 9 months of grinding while working full time at Big Tech and +15 years of experience. After loops with ~30 companies I ended up with about 6 offers.

The offers were all over the shop. Some places downleveled me to mid level while others upleveled me. The comp range was about 3x between the lowest and the highest.

Also I couldn't find much of a pattern on the processes and outcomes. No name companies rejected my application without even a phone screen. Lower tier companies offered me lower levels while famously hard to crack ones offered me higher levels. I also didn't see much correlation on how hard interviews were.

I'm totally convinced the job market is a numbers game and there is a high degree of randomness and luck involved. Discipline and hard work help, but outcomes are far from guaranteed. Last time I almost didn't leave my job since most offers were worse than my then current one, until I got 2 high quality offers at the end.

I have only skimmed through the article but I'm impressed OP got a job as a Research Engineer with no formal background, given that I got doors slammed even after PhD + years of relevant experience. I don't want to diminish the accomplishment, but I would be very wary of anecdata and formulas for success. OP quit his job and focused 100% on the search, and it paid off. I know many more similar stories that didn't pan out, and people had to rush to get lower quality jobs or, in the worst cases, got into long term unemployment.

Just my 2 cents.

Were most interviews focused on DSA and System Design still?

2 years ago, for sure. Every programming interview I had pulled problems from LC. Admittedly it’s quite hard to not, now the database is so big.

Mostly yes. But there were also some take home assignments, some culture fit loops, and many ML rounds, since I was applying for an ML role.

I'm not sure why the OP is using that word: he is working for a (for-profit) company, not at a "lab".

> Mistral AI SAS (French: [mistʁal]) is a French artificial intelligence (AI) company headquartered in Paris.

* https://en.wikipedia.org/wiki/Mistral_AI

I'm not sure why "lab" is being used for all of these companies in this space. Is it an attempt to sweep their money motivations under the rug, or something else?

I don’t think the word lab inherently implies anything about profit motives. Only the focus on frontier research

I'm slightly questioning the term "guide". Basically you have said become a possibly useful researcher in Machine Learning by learning about machine learning, and because Machine Learning takes place on computers it's useful to know how they work too.

This is the equivalent of me saying, if you want to join me researching quantum computing. You need to learn about the quantum and then the computing. Broadly at first then choose a niche.

It's a bit, well obviously.

18 month job hunt is pretty rough, I get that he really wanted this but for most people with "normal" lives outside work this would be a very risky move.

This article is why I see LLM's as a dead end and not real AI. LLM's are somewhat useful for things related to language. LLM's are a decent search engine. But LLM's are not AI.

How would you define AI? It seems they can do a lot of useful things. They are very good for processing language and I find Claude Code to be a better coder than a lot of human programmers.

What I mean to say is that AI encompasses a lot more than just LLM's, and that LLM's in general are a dead-end technology. People think that LLM's are going to keep getting better, but I doubt that they will, they have probably already peaked.

In general I consider neural networks to be a dead-end. But they are still considered to be under the AI umbrella.

The fact that they hired the guy who wrote the article, confirms to me that there is no substance behind LLM technology. It is a waste of time and energy.

Maybe LLM will be to AI what dial up was to the internet. It’s a start that gets things going and attracts investment for better things.

To me, AI is about the ability to solve novel problems. People who think supervised learning can do that usually believe that new knowledge comes from (1) combining old ideas, and (2) trial and error. I don't think that's true. If you think about the state space of all possible ideas, it's clear that humans can explore a very small subset of that space and supervised learning just imitates humans (with great speed and memory). RL also doesn't work, but it at least attempts to learn independently of human knowledge.

That sets the bar pretty high for intelligence and by that standard there probably aren't too many humans who have intelligence.

I think humans and other life forms have this ability on a long enough timeline through evolution.

They’re not AI. But a plural of LLM is LLMs, not LLM’s, unless you’re of course Dutch.

>A learning hack that worked for me was getting YouTube Premium, which allows for downloads and background play, and then listening to these videos while going for a walk

you can also research about youtube-dl/yt-dlp (and using an ad blocker)

The article is somehow even more insufferable than the title.

"Claude, make a me a blog template, that shows how minimalist I am"

tl;dr - the hiring processes in tech get more and more interminable every year.

I am super happy that I am retiring soon. The hiring process these days just sucks.

Reply on news.ycombinator.com