The definition of what constitutes "intuitive" vs "sensor" is itself something we can juggle. A lot of physical scientists would be Keirsey Ns -- they're interested in the unknown, discovery, and aren't conventional, they're intellectual, and so on.
However, many of them would be classified as relatively practical-minded by philosophers, mystics, and so on; speculation is entertained, but what they mean by "possibilities/unknown" is they are curious for new data, new empirical studies. They're not speculative in the sense of an uber-speculative theoretical physicist or philosopher or mystic, i.e. they're too practical to speculate on the "ultimate nature of everything."
Also, that they build theories is explained sufficiently by T, it doesn't really get to N. Technically if we take things to an extreme, anything besides S -- i.e. F, T, N operates on a 'higher plane'. ST science can be a mix of tangible and theorizing. It build theories, but it's not what you'd call very speculative theorizing. It's logical theorizing based mostly on data.
They're still Keirsey Ns, but I think one can reasonably call them sensation types in a more functions-theoretic perspective. The way I frame this isn't too dissimilar to Jung's idea, which seemed to identify natural scientist with a sensation focus, and Ni types with mystics, artists, cranks/prophets, and so on. I just tend to slightly refine his interpretation and include many kinds of philosophical reasoning here, because as much as he might want to place this in T himself, I think the oft-made distinction between the technical and the conceptual/philosophical underscores a T/N difference.