Well, you
can do "data science" in Excel, depending on what you take "data science" to mean. And infrastructure is part of doing data science. I don't think "data science" really means much of anything in practice.
You'll need to carefully define what kind of work you're interested in doing: text mining with NLP, predictive analytics, data mining, ML research, symbolic AI (yeah, that's still around as well), optimization type problems using genetic algorithms, etc. And then probe the company's job posts and dig in with the recruiters, and screen out the bogus stuff.
If somebody advertises for a "data scientist" just treat that as a weak signal that they might be a match, and work from there.
One thought to consider: Instead of working as a FTE in internal IT for some company, you might consider becoming a consultant. You could look at working for a company like Hortonworks or Cloudera, who both have services arms, where you could travel from company to company, working on different kinds of problems.
