Do you think the role data scientist is dead? I keep reading that companies need ML engineers, not data scientists Comments please #datascience #tech
If you provide garbage to ML you'll get garage out.
I mean I interviewed at 10+ companies in the last month for data science roles (not MLE).... so probably not
We have lots of DS at FB still
SQL analysts don’t count
Why not? What is the question here? If there are still a ton of jobs still paying sql analysts to do a/b testing and reporting for $200-300k - does that determine that data science is dead?
Data analyst = Data Scientist i.e. SQL monkey, tableau Data scientist = Applied Scientist/ Reaearch Scientist i.e PhD, heavy on stats
Average company needs probably 5 times less ML people than analytics people... if anything, ml will be automated into irrelevance while analytics and people who actually understand the business and product will continue to be needed
I’ve worked as both SDE and DS. I’d consider the DS work harder. More ambiguity, more required thoughtful communication, and depending on the project, just as much coding. This assumes a baseline ability in programming, which some DS people admittedly lack.
Data Science as an industry is highly unstructured in terms of nomenclature.Most of data science problems are not deterministic unlike SWE. At it's very core- the question is, are you good at problem solving?Would you call a travelling sales man problem a data science problem or a computer science problem?At it's core, there is a significant overlap because fundamentally, you're solving some sort of an optimisation problem. Now, going back to the general pattern of roles. FAANG spends a lot of money in resources in general across roles. The data science actually happens in the AS or RS sort of roles and since they spend a lot of money, they can afford to hire PhDs for it. For most part AS get's paid more than SWEs. Just because some companies are hiring SQL folks and branding it as DS, doesn't make Data Science as a field of study irrelevant to the industry. Whether you like it or not you still need data science to better your recommendation system, better your pricing, make your supply chain efficient etc. None of the above are traditional SWE problems, but then there are significant overlap as well. P.S : Blind is more SWE focused, and there are more SWE than DS/AS in general.
What’s the difference?