Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore how to deploy large language models directly on edge devices in air-gapped environments through this lightning talk from DevConf.US 2025. Learn to build a complete multimodal AI pipeline combining Vision Language Models, Retrieval-Augmented Generation, and LLMs that operates entirely offline using Podman containers on RHEL Edge with GPU CDI on NVIDIA Jetson Orin Nano hardware. Discover the key differences between cloud and edge computing constraints including RAM and power limitations, and examine a container-native architecture designed for low latency, privacy protection, and reproducibility. Watch a pre-recorded demonstration showcasing a camera-to-answer workflow with real device performance metrics including tokens per second and first-token latency measurements. Gain practical insights into deploying rootless, reproducible, air-gapped AI systems using Ramalama for local LLM serving, with operational tips for industrial environments like factories, laboratories, and embedded automotive systems where connectivity is limited or unavailable.
