Stepping Stones to Machine Learning


Seeking opinion on optimal stepping stones for a child to go from Python to Machine Learning.

Summertime, kids out of school, and I'm helping them learn (or should I say keeping them interested and busy so that they don't express Viking genes, sack the house, and wield broadsword-equivalents against each other.)

I discovered CodeCademy and they love it. Son learning Python, daughter HTML. They are cranking through lessons rapidly, and I am very pleased at what they have learned in such a short period of time.
A very smart professor told me that he and his team use Python to code applications that utilize machine learning against medical research and data. I don't see why my kids can't have fun without razing the house, and at the same time learn Python and techniques that might be very useful in a future career (e.g. with medical data)

Naively, I am thinking
Python, then API's, and then experiment with passing info to ML libraries
just to see what they come up with.

As a problem area (not medical data for now) I thought it might be fun to tackle a very simple Augmented Reality app of some sort. If not the end-to-end app, at least a key part of it. A best case would be to include one of my interests. I am intrigued by the idea of taking pictures of Egyptian hieroglyphs, extracting some kind of classification data from it, and then matching it against expert translations/interpretations for that glyph or sequence of glyphs. If an easier problem space makes more sense, then we'll do that. But SOMETHING to keep the kids interested and peaceful. :wink:

Years ago I was in the software industry (coder, RDBMS, business analytics, and more), but I am a novice at ML, and A/R.
I appreciate any advice on
stepping stones from Python to Machine Learning,
and/or A/R