A New Way to Build Edge AI: Agentic Development on NVIDIA Jetson

Building AI at the edge means taking careful coordination across hardware, software, and runtime environments. From selecting compatible JetPack versions and containers to managing memory and configuring inference pipelines, setup and integration can take time away from building the application itself.

Agentic development changes that. By pairing coding agents running with Codex and Claude Code using NVIDIA Jetson Agent Skills, developers can go from a natural-language prompt to a working, optimized edge AI build in far less time. Jetson skills let agents inspect the device, configure the software environment, and construct repeatable AI workflows, reducing the guesswork and iteration that typically slow development down.

In this livestream, we’ll demonstrate how to use a coding agent, such as Codex with the Astra model, in a hands-on Jetson development workflow. You’ll learn how to:

• Install and configure Jetson agent skills.
• Inspect a Jetson device and understand its software and hardware environment.
• Work with JetPack, containers, inference runtimes, and device resources.
• Build and iterate on AI pipeline workflows directly on Jetson.

Join us live, bring your questions, and see how agentic development can help accelerate edge AI development on NVIDIA Jetson—from initial device setup to production-ready AI pipelines