OK, completely out of topic, but it's interesting to notice how broadly they use the term "data science". This goes to confirm my thinking that data science is becoming more and more an "all around data". The various facets of data science I am seeing in this spreadsheet:
- BI
- Data visualization
- Data engineers
- Statistician
- Warehouse
- SQL
- Python-fu
- Marketing
I look at it and I think it's a great prove that data science is not only phd territory, there are so many ways to bring value to the field.
A bit of transparency goes a long way. Where I work (Casetext), our data team does a mix of traditional data engineering, machine learning, graph algorithms, NLP (a mix of off-the-shelf and custom work), information retrieval, customer/market analytics, and experimental design. For us, and probably many other companies in this situation, it helps us to know what a candidate is interested in and good at. It also helps explain what we do to another level of granularity.
Yes, the "data science" label is pretty muddled. It happens (but it is understandable) probably because of the fuzziness of the topics and differing communication styles and goals of engineers, marketers, salespeople, and investors.
I welcome the breadth. Data science is science with data, right? What could be more broad? If anything, it will be nice to break the association somewhat between data science and AI hype.
I also think thats justified. A newer, and more specific term I particularly like is "Machine Learning Engineer", which will probably soon be recoined to "AI Engineer".
We (www.datarevenue.com) basically have to use "AI" now to make it clear what we do. Something that would have made me feel awkward just 3 months ago.
Do you see a substantive difference between AI and ML? "Machine learning" to me is pretty cut and dry, in that anytime something is automated we are employing machine learning, literally teaching a machine to do something. "Artificial intelligence" I have a hard time defining, because I don't have a good definition for "intelligence".
Behind AI I would always expect at least Deep Learning. Machine Learning I use for everything that learns it's own decision boundaries. When it's humans teaching a machine, I'd call it simply automation or expert-system.
Although a lot of "teaching" still goes into feature engineering ...
- BI - Data visualization - Data engineers - Statistician - Warehouse - SQL - Python-fu - Marketing
I look at it and I think it's a great prove that data science is not only phd territory, there are so many ways to bring value to the field.