For developers who are building AI projects, where do you source data for training your AI models? What tools are you using?
It depends entirely on the domain and use case! Wherever data exists for mapping what you have to what you need. For some commonly benchmarked problems like object detection there are massive open-source datasets -- for some less-common problems you'll have to go searching for an application-specific data provider.
One thing you may find useful is going to Google Scholar and looking up your use case -- in particular search for papers that include "benchmark" or "new dataset" to see where academics are posting clean, labelled data they've collected.