Building RAG Agents with LLMs – NVIDIA Deep Learning Institute
Explore NVIDIA’s hands-on workshop to master Retrieval-Augmented Generation (RAG) systems, combining large language models (LLMs) with external knowledge sources for intelligent, context-aware AI agents.
Course Overview
| Aspect | Details |
|---|---|
| Title | Building RAG Agents with LLMs |
| Provider | NVIDIA Deep Learning Institute |
| Format | Instructor-Led Workshop (Live virtual or in-person) |
| Duration | 8 hours |
| Prerequisites | Introductory deep learning knowledge; comfortable with PyTorch; intermediate Python experience |
| Hands-On Environment | GPU-accelerated cloud servers via NVIDIA NGC containers |
| Certificate | NVIDIA DLI Certificate upon completion |
Why Attend?
- Cutting-Edge Techniques: Learn to design AI agents that combine LLM reasoning with real-world data.
- Industry Expertise: Gain insights from NVIDIA engineers and experts in accelerated computing.
- Practical Skills: Build, evaluate, and deploy RAG agents without fine-tuning model weights.
- Career Impact: Elevate your AI development profile with a specialized, in-demand skill set.
Learning Objectives
By the end of this workshop, you will be able to:
- Compose an LLM system that leverages both internal reasoning and external tools for predictable user interactions.
- Design dialog management and document reasoning workflows that maintain context and structure responses.
- Integrate embedding models for efficient similarity searches and guardrailing of AI outputs.
- Implement, modularize, and evaluate a RAG agent capable of answering domain-specific queries without additional fine-tuning.
Workshop Modules
| Module | Topics Covered |
|---|---|
| 1. Introduction | Overview of RAG concepts, use cases, and workshop goals |
| 2. LLM Inference Interfaces | Configuring APIs, batching requests, and latency optimization |
| 3. Pipeline Design | Building workflows with LangChain, Gradio, and LangServe |
| 4. Dialog Management | State maintenance, turn-taking strategies, and structured output formats |
| 5. Document Handling | Loading, parsing, and chunking documents for retrieval |
| 6. Embeddings & Vector Stores | Generating embeddings, setting up vector databases, and similarity search guardrails |
| 7. Evaluation & Q&A | Testing agent performance, benchmarking metrics, and final workshop discussion |
Curriculum sourced from NVIDIA DLI workshop materials.
How to Enroll
- Visit the Course Page
- Register or Log In
Use your NVIDIA account or create one—no prior registration fee. - Select Workshop Dates
Choose a live session that fits your schedule. - Confirm Prerequisites
Ensure foundational deep learning and Python skills. - Join the Session
Attend virtually or in person; complete hands-on labs in the DLI cloud environment.
Usage Tips
- Pre-Workshop Preparation: Brush up on PyTorch basics and familiarize yourself with LangChain.
- Environment Setup: Test your GPU acceleration access on NVIDIA NGC beforehand.
- Active Participation: Engage in live Q&A and use provided datasheets for reference.
- Post-Workshop Practice: Apply your learning by building a simple RAG agent on a personal dataset.
Nvidia Course : Click Here
