Hacker News Reader

Comment by ramraj07 | Hacker News Reader

A good data scientist is a jack of all trades - mediocre programmer, mediocre ML modeler, mediocre ETL architect, mediocre statistician and a mediocre analyst. You hire this person because you don't need a specialist in each of these fields But you need someone who can do these things. Also you don't want to pay a metric ton of money so you can't expect some superstar.

View on HN

Good data scientists also need to be great at a couple of things. They must be excellent presenters, and excellent diplomats. Getting the data is the hardest part. Data silos are create all over the place and getting access to that data requires speaking to a lot of higher ups.

Can confirm. Am data scientist. Am mediocre at everything :)

> Also you don't want to pay a metric ton of money

Doesn't the job title of "data scientist" still correlate with high pay, or has the hype started to wane?

Depends on the location and company. In a smaller city working for a smaller company an entry level data scientist might be happy to pull in 60k. In a major city working for a large company you can at the very least double that number.

If the person can do all of these in a way that makes them a profitable investment, it doesn't really seem so high

A data scientist doesn't command the same premium as a machine learning engineer or a kubernetes expert, I'm afraid

Depends where. I'm at Amazon, and they are really serious about not letting DS become just business analysts who just SQL/Python. But elsewhere I do get the impression there is some dilution. Although honestly, that's natural. Just like how in the SDE world there are template web devs and hardcore infrastructure guys, who get paid massively different amounts but have the same job title, that's happening to DS.

These people often have PhD's, I doubt they are mediocre statisticians or mediocre analysts.

Hmmmm. I wish your implication was correct, but I fear it's not.

Reply on news.ycombinator.com