@JohnMufasa1@GergelyOrosz@peekknuf Data labellers play a part too. But the best data scientists and ML engineers label data. It's good practice and should be expected of the role.
@GergelyOrosz@peekknuf Data labelling is the highest ROI activity for anyone developing ML systems, esp early on in the projects/products life. If you're not labelling data or interrogating your dataset to the point you know/remember many specific examples, you're not going to build something useful.
@lennysan Whole classes of data science projects are gone. LLMs are so versatile, you can apply them in many cases where before a whole DS project was required.
Predictive modelling you still need a data scientist but even now one great one gets so much leverage, team sizes can shrink
@Sam__Enright@o_mcpartlin@_fitzwilliam@cvi94 I run DS teams in Dublin and London and worked in DL research team in multinat previously.
Actual legit ML talent in Dublin is incredibly low. Majority of people are CV stuffing with Coursera projects.
Very few would make it in a top ML team/company. Better talent in London.
@Keyes Or tax treatment of share options when exercised. You either pay CGT on unrealized gains if exercised, or if a part of a liquidation event, you pay income tax at 52%. It is insanity