12.00 – 12.25
Artificial Intelligence (AI) and Machine Learning (ML) workloads rely on efficient and scalable System on Module architectures. These architectures must support current performance requirements, remain adaptable for future application demands and enable sustainable long-term lifecycles for deployed devices.
In this session, we explore how integrated Neural Processing Units (NPUs) in the i.MX95 processor module enable reliable and power-efficient on-device processing for different types of AI and ML tasks. We also cover the practical workflow for adding these capabilities to embedded designs, including key design considerations and real-world examples.
Finally, we demonstrate an application-oriented approach to SoM selection with AI and ML workflow integration. This approach enables more capable edge devices, reduces cloud dependency, and provides a clear path to implementing algorithms on the hardware.
Speaker: Pierluigi Passaro – Variscite (on behalf of Batenburg Applied Technologies)