Hi Renato, wish I could provide a better solution but scraping on LinkedIn isn't quite feasible at large scale. I tried the approach of only run the scrape script when ever a user makes request but at thousands concurrent users, it would always fail. I even tried putting them all in queues but it still failed for large number of user.

It is entirely possible as long as the script isn't running as scraper (I know, shocking!). I had a stupid idea that the script would run like a user, open the Linkedin Profiles, take a screenshot then have another OCR processor to extract the text, then another text processor (may be NLP) to put the texts in the right places. But that's just way too complicated.