I've worked in this business. Non-invasive blood glucose monitoring has been a holy grail of analytical science for 50+ years. I've known two people who have gone bankrupt trying to do it.

The spectrum contains signatures of hundreds of compounds. All of the variation has to be accounted for. It's actually a ridiculously hard problem.

Still, holy grails being what they are, I'm still excited when I see that someone is giving it another chance.