Nvidia Courses

Building an AI-Enabled Network Video Recorder with NVIDIA Metropolis Microservices for Jetson

Learn to build an AI-enabled Network Video Recorder using NVIDIA Metropolis microservices for Jetson. Covers system setup, microservices, testing, diagnostics, and code samples for various operations in edge AI applications.

• YouTube
• 24 minutes
• Self-Paced
• Free Video

Step-by-Step Installation of NVIDIA Virtual GPU Software on VMware vSphere

Install and configure NVIDIA Virtual GPU software on VMware vSphere for graphics, VDI, or AI workloads. Learn server preparation, GPU Manager installation, vCenter setup, guest driver installation, and VM licensing.

• YouTube
• 36 minutes
• Self-Paced
• Free Video

DeepSeek R1 Performance Optimization to Push the Throughput Performance Boundary

Discover how to optimize DeepSeek R1 throughput scenarios and achieve world-class GPU inference performance using NVIDIA Blackwell architecture from industry experts.

• YouTube
• 1 hour 3 minutes
• Self-Paced
• Free Video

The Practice of Doing Performance Analysis and Optimization with TensorRT-LLM

Master TensorRT-LLM performance analysis and optimization techniques using expert tools, profiling methods, and bottleneck identification strategies.

• YouTube
• 54 minutes
• Self-Paced
• Free Video

DGX Spark - Accelerated cuPyNumeric on your Desktop Cluster

Discover how to scale Python and NumPy programs across multi-GPU systems using NVIDIA cuPyNumeric without complex parallel computing code on DGX Spark desktop clusters.

• YouTube
• 33 minutes
• Self-Paced
• Free Video

Learn With Me - Building Grounded AI Agents with Nemotron and RAG

Discover how to build complete AI agents from scratch using Nemotron models and RAG systems in this hands-on session with step-by-step implementation guidance.

• YouTube
• 48 minutes
• Self-Paced
• Free Video

Build a Bash Computer Operator Agent with NVIDIA Nemotron

Discover how to build a natural-language Bash agent using NVIDIA Nemotron in under an hour with just 200 lines of Python code and minimal dependencies.

• YouTube
• 47 minutes
• Self-Paced
• Free Video

DGX Spark - Powering EPRI's AI Research

Discover how EPRI leverages NVIDIA DGX Spark to run advanced open-weight models like gpt-oss-120b for secure, AI-powered document research in energy innovation.

• YouTube
• 31 minutes
• Self-Paced
• Free Video

Accelerate Robotics and Real-Time AI Inference on NVIDIA Jetson Thor

Unlock the power of NVIDIA Jetson Thor for next-generation robotics with 3.5X faster generative AI performance and real-time reasoning capabilities for physical AI applications.

• YouTube
• 1 hour 2 minutes
• Self-Paced
• Free Video

AI Performance Benchmarking - Dynamo and Other LLM Endpoints

Discover NVIDIA Dynamo's LLM acceleration features and AIPerf benchmarking tool through hands-on demos, covering PyTorch, TensorRT-LLM, and vLLM integration for optimized AI inference.

• YouTube
• 43 minutes
• Self-Paced
• Free Video

Accelerating Science and Engineering With NVIDIA CUDA-X Libraries

Explore how NVIDIA's 900+ CUDA-X libraries revolutionize industries from data science to quantum computing, driving breakthrough scientific and engineering innovations.

• YouTube
• 32 minutes
• Self-Paced
• Free Video

CUDA 13.0 - New Features and Beyond

Discover CUDA 13.0's latest features and future GPU computing developments in this engineering-focused talk by NVIDIA's technical product lead.

• YouTube
• 42 minutes
• Self-Paced
• Free Video

Open-Source AI 101 - Enabling American Innovation

Discover how open-source AI democratizes innovation, drives economic growth, and enables transparent collaboration through expert insights from NVIDIA, OSS Capital, and academia.

• YouTube
• 42 minutes
• Self-Paced
• Free Video

Scaling Python Analytics - NVIDIA cuPyNumeric and Legate Boost for HPC

Discover how to scale Python data analytics from laptop to supercomputer using NVIDIA cuPyNumeric and Legate Boost for distributed computing across multiple nodes and GPUs.

• YouTube
• 28 minutes
• Self-Paced
• Free Video

Deep Dive - How to Use cuTile Python

Dive into cuTile Python for tile-based parallel programming with CUDA 13.1's virtual instruction set, abstracting tensor core complexities for high-level algorithm development.

• YouTube
• 16 minutes
• Self-Paced
• Free Video