nemo-guardrails

This project applies the core knowledge from the LLMOps module, including the design and implementation of the API Layer, Inference Layer, Observability Layer, Cache Layer, Guardrails Layer, Routing Layer, and the Data Ingestion Pipeline.

LLM tutorial materials include but not limited to NVIDIA NeMo, TensorRT-LLM, Triton Inference Server, and NeMo Guardrails.

💂🏼 Build your Documentation AI with Nemo Guardrails

🔥 QFIRE — a prompt firewall for LLM applications (proxy + CLI + benchmark harness) in Rust (100+ firewall rules).

Responsible AI Masterclass (June 2024 Run)

An Agentic RAG-powered chatbot delivering intelligent, fact-grounded and context-aware answers for the DIEM Department at the University of Salerno.

RAG with Redis. Parallel implementations with/without guardrails.

A simple Retrieval-Augmented Generation (RAG) pipeline for answering questions based on website content. This project combines retrieval of relevant website information with generative models to deliver contextually accurate answers.

An end-to-end AI pipeline that scrapes LinkedIn profile and post data to predict 16-personality (MBTI) types using a RAG-enhanced LLM fine-tuned with LoRA. It automates data collection, preprocessing, storage in PostgreSQL, and personality inference from real-world behavioral and linguistic patterns.

This bootcamp is designed to give NLP researchers an end-to-end overview on the fundamentals of NVIDIA NeMo framework, complete solution for building large language models. It will also have hands-on exercises complimented by tutorials, code snippets, and presentations to help researchers kick-start with NeMo LLM Service and Guardrails.

GitDoc is your ultimate GitHub Documentation Explorer! It's your trusty sidekick for navigating through the vast world of open-source projects, making code exploration and documentation retrieval a breeze. 🚀

Hands-on AI security research covering prompt injection, LLM red-teaming, AI supply chain risks, and data poisoning. Features a Groq-powered interactive simulator (16 scenarios, 5 live LLM chats) and a NeMo Guardrails + Colang production demo. All 25 TryHackMe AI1 rooms mapped to OWASP LLM Top 10 and MITRE ATLAS.

Dive into the world of LLM Guardrails using tools like NVIDIA’s NeMo Guardrails. Discover the mechanisms that ensure applications produce reliable, robust, safe, and ethical outputs, and understand their crucial role in LLMs

Real-time intent preservation layer for LLM agents. Stops hallucination drift before it executes.

Hands-on examples for NVIDIA NeMo Guardrails. Learn to add safety guardrails to LLM applications through executable Jupyter notebooks paired with a technical article.

Generative AI RAG Chatbot for Electricity and Gas Company

Production-grade, multi-tenant Agentic RAG platform featuring LangGraph CRAG orchestration, hybrid dense/sparse vector retrieval via Qdrant, RBAC isolation, Redis caching, and automated CI/CD security gating.

Multi-layered AI-powered Web Application Firewall for LLMs — fast regex triage, NeMo intent classification, CAMEL multi-agent verification, and adaptive feedback loop to block prompt injection, jailbreaks, and adversarial attacks with explainable decisions.

Production-grade Agentic RAG: LangGraph planner/retriever/responder, NeMo Guardrails safety gate, Portkey LLM gateway (Groq Llama 3.3), Qdrant + FlashRank two-stage retrieval, Gemini embeddings, RAGAS evals, Logfire + LangSmith observability.

Hands-on demos for the Pluralsight course: Generative AI Data Privacy and Safe Use for Developers. Covers PII masking, prompt injection attacks and defenses, five guardrail rail types (input, retrieval, dialog, execution, output), evaluation release gates, and dashboards mapped to EU AI Act, NIST AI RMF, and ISO/IEC 42001.