What I did was to use ai to solve a problem in one of those upcoming areas I thought would become important (crypto currency trading in my case), I created a fo

What I did was to use ai to solve a problem in one of those upcoming areas I thought would become important (crypto currency trading in my case), I created a forecasting/ai prediction system, with a stats dashboard https://bitbank.nz https://bitbank.nz, I charge a subscription fee and am transparent about the accuracy over time to build a long term loyalty. This strategy worked well for me, use standard machine learning techniques to solve an existing problem, make the system as easy to interact with as possible, UI, API ect. Try to gather the customers data for them because a lot of machine learning is getting a good size dataset and aggressively training a flexible model on it. a good dataset can reduce friction if your customers can simply sign up and see value straight away. I'd say (biased) do what I did... major in compsci which keeps things flexible, do your own research and create your own company to break into the ai/machine learning space which can be hard to get practical experience in for a new grad