return to speakers list
Building AI models that run well in resource-constrained MCUs is extremely challenging. As with any AI model, collecting and labeling quality datasets can be time-consuming, expensive, and open-ended. From there, building a compact inference model that is both accurate and fits within the microcontroller has traditionally required hand-crafted code from data scientists. Learn how this process can be streamlined, made simple, and targeted to even the smallest 8-bit microcontrollers.