Free NVIDIA courses and the skills you'll build
Every course here is free to start. Each one is matched to the day of the 30-day plan where you'll use it, so you learn a thing in the morning and build it on your Jetson the same afternoon.
- 15
- free courses
- 10
- from NVIDIA DLI and Jetson AI Lab
- 12
- skills by Day 30
- $0
- to follow the whole path
Skills you'll have by Day 30
GPU and CUDA basics
Kernels, memory, power modes, and reading jtop.
Days 1–2PyTorch training
Data splits, training loops, loss curves, overfitting and how to fix it.
Days 3–5Computer vision
Transfer learning, augmentation, confusion matrices on your own photos.
Days 4–5Edge model export
ONNX, TensorRT engines and TFLite int8 with accuracy and latency checks.
Days 6, 17Time-series and Transformers
Sensor windows, 1D-CNNs and a tiny Transformer compared fairly.
Days 8–10Audio DSP
MFCC from scratch, golden vectors and audio classifiers.
Days 11–12Object detection
Labelling, a centroid detector, F1 and overlays.
Days 15–17On-device LLMs
llama.cpp with CUDA, Nemotron 3 Nano 4B, tokens/s and memory.
Day 19RAG and vision-language helpers
Embeddings, cited answers, Cosmos Reason 2 and NanoOWL pre-labelling.
Days 19–20ML engineering
Config-driven pipelines, packages, tests and reproducible runs.
Days 13, 22AI APIs and web apps
FastAPI workers with live progress and a Next.js front end.
Days 23–27Agents and MCP (next)
Tool-calling agents and MCP servers for Month 2.
Days 29–30Free courses
Getting Started with AI on Jetson Nano
NVIDIA DLI · Set up a Jetson, collect images and train a classifier on the device. The concepts carry straight over to your Orin NX.
An Even Easier Introduction to CUDA
NVIDIA DLI · Write and launch your first CUDA kernels and see why the GPU wins on parallel work.
Building A Brain in 10 Minutes!
NVIDIA DLI · The fastest way to see how a neural network learns, before you train your own.
Generative AI Explained
NVIDIA DLI · What generative models and transformers are, and where they fit. Good background for the Day 10 Transformer.
Building Video AI Applications at the Edge on Jetson Nano
NVIDIA DLI · Build video analytics pipelines on a Jetson with DeepStream — the production path for your detector.
Augmenting LLMs using Retrieval Augmented Generation
NVIDIA DLI · A one-hour primer on RAG: retrieve first, then answer only from what you found.
Building RAG Agents for LLMs
NVIDIA DLI · Go from RAG to agents that plan, call tools and cite sources — the path to the SutraAI assistant.
Accelerate Data Science Workflows with Zero Code Changes
NVIDIA DLI · Speed up pandas-style data prep on the GPU with RAPIDS, without rewriting your code.
Jetson AI Lab tutorials
NVIDIA · Exact commands for running LLMs, VLMs and NanoOWL on Jetson boards like yours.
NVIDIA Agentic AI Learning Path
NVIDIA · NVIDIA's curated path for building AI agents with Nemotron and NIM.
learnpytorch.io — PyTorch for Deep Learning
Zero to Mastery · The beginner PyTorch path behind Weeks 1–2: tensors, training loops, vision, custom data, transfer learning.
The Sound of AI
Valerio Velardo · Audio signal processing for ML, explained step by step: spectrograms, MFCC and audio classifiers.
Hugging Face MCP Course
Hugging Face · Learn the Model Context Protocol and build MCP servers so assistants can call your Jetson tools.
FastAPI official tutorial
FastAPI · Everything Step 15 needs: routes, uploads, background work and streaming responses.
Next.js Learn
Vercel · The official Next.js course, for the upload → train → download slice in Step 16.
How to get the most from free courses
- Create a free NVIDIA account and join the NVIDIA Developer Program; it also includes a complimentary self-paced DLI course.
- Do the course on the morning of its day, then build the matching step on your Jetson in the afternoon.
- Write three notes in your own words in
BUILDLOG.md. Teaching it back is what makes it stick. - Some DLI courses issue a certificate when you pass their assessment. Check each course page and add the ones you earn to your profile.
Prices and free status are set by each provider and can change. The NVIDIA courses above were listed as free on NVIDIA's self-paced training page in October 2026. Paid options (Udemy and paid DLI) are on the courses and models page.