Hi there, my approach here will be reverse engineering. Look for the most interesting papers in the area do you want to gain experience. Try to reproduce the experiments in details, document your results and even try to make new experiments to see if you find something new. Ask questions to the authors about your problems. I'm sure that making this in your after ours time, you will gain a good intuition of what research means and you will have some learning lessons to share with others.
I am more or less in the same position, I'm 39 old with a very good programmer job, in this case with a Java stack. My personal opinion is that Java, J2EE, .Net will continue to be relevant because many business have migrated his legacy cobol apps to this stack, but sad but true, .Net and Java are the new cobol... you can retired woking with they, but they are not the future any more.
The future is in languages as Scala, Python, Ruby, Javascript and ideas as Node.js, NoSQL.... you want to create a new stack to play... review that are using the top startup and you will see a clear map.
My strategy is to use this new technologies in my side projects, nobody pay me for that (yet) but allow me mastering what is comming. Make this transition with a full time job, family, wife and kids is a big challenge. I want be part of the new wave in IT and no spend the rest of my life in legacy technologies.
Let me know and I can show you more ideas how start the trip...