If Linus Torvalds decided to pick up data engineering tomorrow, he’d be far better at it in 3 months than the average dev who’s been in the field for 10 years.

When you filter on years of experience, you’re inadvertently biasing towards less intelligent candidates. Smart people with that much experience will have enough mastery that they’re not looking for a mid level position. Polymaths who jump between fields, quickly building up competence won’t meet your arbitrary threshold.

bold claim, I'd want to take a bet with you for real money on that

decade ago was 2011, wasn't it time where applied data engineering started becoming cool? thus majority of people shouldnt be those, who started learing it due to $$, but out of curiosity?

idk

From wiki

"data engineering, is a software engineering approach to designing and developing information systems."

"An information system (IS) is a formal, sociotechnical, organizational system designed to collect, process, store, and distribute information."

So what do you think you would you call it?

>"data engineering, is a software engineering approach to designing and developing information systems."

this description is so broad, that every SE would be data engineer now?

for me git is just text comparison, management + some kind of visualisation tool (so it kinda sounds like data X stuff, doesn't it? I'll back to it below)

meanwhile data X jobs seems to be more focused on:

understanding business domain

using e.g statistical methods

working with huge data collections with various tools like hadoop, spark and fancy keywords like data lake, warehouse

>Data science is an interdisciplinary field that uses scientific methods, processes, algorithms and systems to extract knowledge and insights from structured and unstructured data,[1][2] and apply knowledge and actionable insights from data across a broad range of application domains. Data science is related to data mining, machine learning and big data.

>Data science is a "concept to unify statistics, data analysis, informatics, and their related methods" in order to "understand and analyze actual phenomena" with data. It uses techniques and theories drawn from many fields within the context of mathematics, statistics, computer science, information science, and domain knowledge.

I feel like git doesn't fits this description, I'd say it's more like handy tool that handles snapshoots of some text

_______

or maybe I'm misunderstanding relation between "data engineer" and "data science"?

That's not just what git is.

> meanwhile data X jobs seems to be more focused on:

> understanding business domain

git is absolutely about that. The technical aspects of git (which themselves are revolutionary data engineering innovations) are entirely driven by the problem domain. The people and their processes. It wasn't just developed out of thin air and people by pure coincidence decided they liked it and started using it!

> working with huge data collections with various tools like hadoop, spark and fancy keywords like data lake, warehouse [etc]

That's not an adequate definition of data engineer.

[data science]

Yeah we're talking about data engineer. Although you could say there was most probably data science work behind git to find performance and storage requirements and optimize it, just not so much of the buzzword type of data science.

...and lots of people don't, even after 10 years working in the field. Not because they're not smart, of course, but it's common.

It's not like they are asking for more than 8 years of experience in Pytorch.

But the position is for working on graphics rendering, a completely unrelated field. Are you supposed to have had two careers at the same time, working eight years in AI while also working all that same time on graphics?

Since Facebook is doing VR, and this position is about VR, I can easily see the potential of using ML to make the most of sketchy/low-quality models and textures people might want to use in worlds. Imagine if you could upload a rough model, describe what you want, and get it made for you.

I mean obviously a lot of people work in that area but the overlap with AI/ML simultaneously seems very small, for 8+ years in one person.

And of course if you're in research then it's hard to tick boxes like "leading consumer-based projects with industry-wide impact" and "mentoring/influencing senior engineers across organizations" and so on.

The number of people working in "those areas," defined narrowly, is much bigger than it was eight years ago. That makes this a big requirement. People who are skilled, deep or enthusiastic enough to make that 8 years experience better than 4 are not likely to just show up, while also ticking the other boxes. Narrowing the pool past a certain point is a different type of problem.

In any case, these aren't to be taken to literally.

[1] https://en.wikipedia.org/wiki/Online_analytical_processing

There's lots of things invented decades ago that would be a parlor trick today.

Facebook pays more than enough to get world class experts so why would they not aim to hire them?