TinyML, the ability to run machine learning on small, cheap, battery-powered devices, makes a lot of new applications possible, including uses in medicine, agriculture, consumer electronics, wildlife conservation, and many other areas. This new technology seems like it could be especially suited to deployment in developing countries since the hardware itself is not costly, it can work in places with little or no data coverage, and offers the chance to adapt to local conditions and languages. It’s also such a new area that there are significant opportunities for researchers and startups to have a lot of impact without requiring a lot of capital. In this talk, I’ll talk about the underlying field, share some ideas on how I think it might be applicable in Kenya, and hopefully learn from the audience about what challenges you think are most relevant and interesting.