Let's assume contrary to @shane that you have correctly expressed your real motivations.
If your motivation is to be 'avant-garde' or to be contrarian, then you have a problem. Not everyone can be contrarian, and purposely flout the opinions of the majority for sport. Nor can everyone simultaneously 'push the boundaries' on purpose by constantly undermining popular opinion. We value those actions in a modern competitive society because they keep things evolving. But it seems clear that Kant would not.
To play with expectations, with those motivations, is not universalizable. Knowingly disturbing people only for the sake of your own ego, or to improve some abstract process, is treating them as mere means. Especially because the people you are most likely to disturb are those who will not value the disruption, even if it does, in the end, improve their lives, or at least those of their progeny.
If this is not a chosen motivation, and you like what you like in all honesty, there is no problem with being out of step by simply being yourself. Sentiments are not moral, according to Kant, so you are free to have what sentiments you have, as long as your motives and considerations are right.
--- Here is a whole different framing for this argument, focussed on the questions raised in the comments. It is the same argument, just much, much pickier.
It is called the categorical imperative to rule out arguments tied to contingencies. Theoretically, for Kant, every truly moral argument should be understood by any adequate intelligence, not just other humans, and he tried to analyze out the kinds of bases that an intelligence would have to have. If the intention cannot be put in terms of a certain sort, it is not 'categorical' and cannot describe a worthy maxim because it could not be translated into a usable form for a different kind of intelligence that recognized a different range of contingency.
There are huge holes in his analysis of what is and is not a category, or how exactly categories should be combined into a maxim if other intelligences do not necessarily have languages. It is one of the primary weaknesses of the resulting ethics. But there is an idea that we should be able to develop an intuition for what is and is not adequately general, even if Kant's attempts to do so are either too hard to understand or too incompletely expressed for most of us to use them.
It is clear that sentiments don't qualify, so yearnings, insecurities or other motivations for winning can't be an important part of a moral argument. Neither can natural human desire for a given outcome if it is just an emotional position contingent upon our humanity.
But 'beauty' does qualify. Doing something with others well is a form of beauty, even if that thing is competition. So it is probably universalizable to say that if you join in a communal endeavor, you should intend to do as well as possible. Everyone can apply that to their own behavior. So to the degree that choosing beauty is done competitively, we are all still fine with most forms of competition.
Cultivating and recognizing originality could also be an important aspect of beauty, especially if variety multiplies the opportunities for beauty itself. But in this case, the approach to finding originality stems from rebelling against the status quo. Were it universalized, the status quo against which one is rebelling would cease to exist. Like 'truth' in the classic Kantian argument against lying, defying standards dissolves the opportunity for standards to exist long enough to be defied. So we would lose the ability to follow our maxim if we agreed upon it.