Just as nobody wanted to write the IP address, why is still "normal" to dial a completely random number in the telephone to call someone. Why is the overhead of similar structure as we have on internet not applicable in telephones.
The advantage of using the same system is a similar API for users, and the opportunity to use identical query language to run both matching queries and analyses like this one in cuGraph. For anyone using graph DBs, this might be relevant, while on the other side, achieving this in a few lines of Python seems like a good deal for typical analysts like you mentioned.
I've been thinking about developing a tool that would help to transform the graph DB data into a Pytorch dataset and use it with a custom model. Are there some red flags that I should have in mind regarding this tight coupling between ML and DB?
What are some examples of products or tools which deliver an ML experience by plug & play approach? Are there any databases or message queues that easily connect to the ML inference model which uses their data?
Working in R&D can sometimes be a pain in the a*, especially when designing a process that would fit the whole team. I've heard about and have worked on projects where there were no daily meetings, no discussion about technology across the team, and it was kind of free-form till the due date. Enforcing rules on data scientists often backfires with their productivity too.How do you tackle...