This Developer Day was given in March 2025 during NVIDIA GTC.
AI Adoption is transforming industries, organizations and daily operations – learn how to bring security fundamentals to the next generation of agentic AI application and deploy with confidence.
In this developer day, learn how to break down an agentic AI workflow, the AI Virtual Assistant NVIDIA Blueprint into its core components and:
- Analyze the blueprint for sample attacks and threats
- Identify general security mitigations
- Discover weakness in the LLM model with
garak - Apply guardrails to mitigate LLM-specific weaknesses with NeMo Guardrails
Access the notebooks through NVIDIA Brev:
The developer day contains several Jupyter notebooks that demostrate the usage of garak and NeMo Guardrails to discover model weakness and apply mitigations.
Follow this order for notebooks:
setup.ipynb: Initial setup and configuration of API tokensgarak_demo.ipynb: Demonstration of Garak for LLM security testingguardrails_demo.ipynb: Demonstration of NeMo Guardrails implementationguardrails_garak_demo.ipynb: Combined demo of Guardrails and Garak
NVIDIA AI Blueprints are reference examples that illustrate how NVIDIA NIM and NVIDIA AI Enterprise software can be leveraged to build innovative solutions
NVIDIA NIM is a set of accelerated inference microservices that allow organizations to run AI models on NVIDIA GPUs anywhere.
NVIDIA Brev provides streamlined access to NVIDIA GPU instances on popular cloud platforms, automatic environment setup, and flexible deployment options, enabling developers to start experimenting instantly.
garak helps developers discover weaknesses and unwanted behaviors in anything using language model technology.
NeMo Guardrails is an open-source toolkit for easily adding programmable guardrails to LLM-based conversational applications.
