Jetson-Accelerated Edge AI:
The Value of SLMs and the Software That Enables Them
This talk introduces why edge AI is essential for accelerating on-site decision-making, the utility of SLMs, and the Jetson platform and related software as the execution foundation. Edge AI delivers low latency, offline resilience, and data sovereignty, avoiding constraints from network quality and transfer costs. SLMs operate under memory and power limits with short inference latency for real-time processing. Jetson achieves high perf/W and low token latency with its GPU and rich I/O. We will also introduce the Jetson family and the software that supports generative AI on Jetson.
- Generative AI
Speaker
NVIDIA
Solution Architecture & Engineering
Senior Solution Architect
Masaya Ogushi
As a Senior Solutions Architect at NVIDIA, I provide technical support to customers, particularly in the fields of robotics and factory automation. I also act as an evangelist, delivering talks and leading hands-on workshops. Before joining NVIDIA, I worked at a systems integrator developing financial systems, and at a startup where I built a 3D object detection engine and time-series anomaly detection systems.