Thanks for your comments.
I agree that Data Science and Software engineering are becoming integrated. I don't doubt it.
In fact I myself have enough software engineering chops myself (but not to the level of an expert programmer) to make my Data Science solutions into MVPs.
Here is some sample of my work
2.https://towardsdatascience.com/how-to-host-a-r-shiny-app-on-aws-cloud-in-7-simple-steps-5595e7885722
3.https://towardsdatascience.com/how-to-dockerize-an-r-shiny-app-part-1-d4267659312a
4.https://towardsdatascience.com/spacy-redis-magic-60f25c21303d
Whenever my nature of job tilted towards being more software engineer-ish, I found that I did not have enough time to hone my core data science skills (statistics, maths and ML algorithms).
The whole point of the article was, one can't apply software engineering practices and jargons to Data Science. Data Science in itself is a huge field, to do justice to the job one has to know the deep workings of algorithms, understand the maths/stats behind it.
With limited time, surely one can't master Data Science and be great at UI design and be great at database architecture/backend coding.
I have also come across people who are solely solicited because of their data science prowess not because they know a little bit of everything, Expertise is still respected. If expertise is not respected, then the data science as a field needs to mature.
The branding of 'Full stack Data Scientist' is more for cutting costs than any other reason. Make 1 person do a team's job while paying him/her 1 person's salary.
Many companies are realizing that Data Science solutions / products are failing despite the good 'data engineering'. The problem is clearly at the level of not knowing Data Science deeply. For everything, naive data scientists throws Random forest or GBM or any off the shelf algorithms
Accuracy is not the only metric to go by, one also needs to have the knowledge when things will fail and what to do when things fail. Sadly, not many data scientist today know what to do when things fail or how to fix it. Read this good article if you get time .
https://a16z.com/2020/08/12/taming-the-tail-adventures-in-improving-ai-economics/
