Ask HN: What ML problems do you have? looking for startups that need ML help

I'm part of a team of ex-Google/Microsoft/MIT PhDs working on a "Stripe for Machine Learning" startup (meaning: building products that allow companies to instantly apply ML to their business, a lot easier than all the other workflows out there). We're trying to better understand different needs and applications, so I'm wondering: what kinds of problems do you face where machine learning might potentially benefit, and what troubles do you encounter?

For example: are you having trouble with spam or categorizing your users/items? Do you wish you had better search or recommendations? Is it too expensive or time-consuming to find machine learning help yourself?

We're based in SF, so we'd also love to meet up with anyone in the Bay Area who wants to chat in person. (And we're happy of course to chat/email/VC with anyone anywhere else!) Just hit me up over email (in my profile).

(Also, our first products are actually out and ready -- you could get up and running with problems like spam detection, user/item categorization, sentiment analysis, webpage classification, etc in literally five minutes and two lines of code. We're not going with a full launch yet because we want to make sure we can 100% support all our early customers, but if you're interested in being part of our private launch or just curious to learn more, definitely reach out!)