[noob question] I have a natural language project I want to work up to. No ML background, but majored in math. What is the path from here to there?

Tech skills include object-oriented JS, SQL, xml, and common related skills for each, like JSON, database design, web services, etc. It's my impression, however, that much of this is useless here. True/false?

I was thinking, generally, of tackling typical beginner ML projects like character recognition which have been tried and tested enough that reference material should be easy to find, then using my experiences there as a stepping stone to more complex problems.

What do you guys think?

Thanks for your time.