How Oracle Helps Secure Containerized LLMs on Kubernetes with NVIDIA

Combining application-level guardrails with container-level security measures creates a robust, multilayered defense for LLMs. As AI solutions become more ubiquitous, developers are concerned about security at the front door – guarding against prompt injections, and other threats, is now of vital importance.

Join this session to learn how you can use Oracle Cloud Infrastructure Kubernetes Engine (OKE) and NVIDIA NeMo™ Guardrails to protect user-facing AI solutions that include prompts of almost any kind. You’ll see how we set up OKE and NeMo Guardrails to find the right balance between UX and security. These principles can be used in other contexts, but the focus will be on protecting inputs from breaching security controls that would otherwise have been in place with previous input methods that were not tied to LLMs.

Developers will be able to implement this knowledge on their own projects, setting up guardrails for prompts to avoid problems with LLMs and AI projects.


Duration: 1 hour