I would say it depends on whether you enjoy the data engineering part of data science. E.g. writing SQL queries, writing scripts to scrape data from APIs, tracking down edge cases and verifying correctness, designing clean and scalable pipelines, writing scripts to make messy data cleaner, automating report generation, data warehousing. This is very detail-oriented work and requires a certain kind of personality to excel at (imo). But if you build this skill set there is extreme demand for you that will be timeless. This is also the foundation for all data science.

If what you have in mind is feeding data into TensorFlow or PyTorch, and tuning parameters to get some magic out the other end, I don't think there's as much of a market (there are many juniors looking for work from bootcamp programs), and you risk yet another "AI Winter" backlash making companies go sour on these sorts of approaches.

I think you could also easily getting a job as a master of a particular analytics tool, like MixPanel or Google Analytics. But I'm not sure how timeless this will be: in my experience sooner rather than later you'll want deeper and more customized tools and run into the limitations of simple event-based analytics services.

It depends on where you want to start out, or go. Most developers of all types will wind up working in a dozen different environments and tools that weren't even a thing when starting.

In the end, you will never be able to stop and rest on your laurels, as you will be unable to find work if you do. Constantly learn, constantly grow, and when starting out, go above and beyond any assigned work.