Broadening access to both computational and educational resources is criticalto diffusing machine-learning (ML) innovation. However, today, most MLresources and experts are siloed in a few countries and organizations. In thispaper, we describe our pedagogical approach to increasing access to applied MLthrough a massive open online course (MOOC) on Tiny Machine Learning (TinyML).We suggest that TinyML, ML on resource-constrained embedded devices, is anattractive means to widen access because TinyML both leverages low-cost andglobally accessible hardware, and encourages the development of complete,self-contained applications, from data collection to deployment. To this end, acollaboration between academia (Harvard University) and industry (Google)produced a four-part MOOC that provides application-oriented instruction on howto develop solutions using TinyML. The series is openly available on the edXMOOC platform, has no prerequisites beyond basic programming, and is designedfor learners from a global variety of backgrounds. It introduces pupils toreal-world applications, ML algorithms, data-set engineering, and the ethicalconsiderations of these technologies via hands-on programming and deployment ofTinyML applications in both the cloud and their own microcontrollers. Tofacilitate continued learning, community building, and collaboration beyond thecourses, we launched a standalone website, a forum, a chat, and an optionalcourse-project competition. We also released the course materials publicly,hoping they will inspire the next generation of ML practitioners and educatorsand further broaden access to cutting-edge ML technologies.