Deep Learning Institute · self-paced

NVIDIA's Free AI Courses

The genuinely-free, in-browser courses worth your time — grouped by topic, with a learning path built for someone who ships Python and builds LLM apps.

Free NVIDIA account Runs in-browser (GPU cloud) No local install Badge / certificate on select courses

This is a July 2026 snapshot. NVIDIA has no stable "all free" page and rotates free/paid status. Always confirm against the live Price → Free filter, and search by title — course IDs rotate.

Your path — GenAI & RAG, fastest useful route

~4 hours (steps 1–4) gets you the mental model; step 5 is the hands-on payoff whose stack mirrors a real Flask + LLM app.

1

Building a Brain in 10 Minutes

Trivial warm-up — confirms the notebook environment works, gives neural-net intuition.

~10 min · Beginner

2

Generative AI Explained

Shared vocabulary — transformers, diffusion. Skip if you already know it.

~2 hr · Beginner · no code

3

Augment your LLM Using RAG

The concept of retrieval-augmented generation before you build it.

~1 hr · Beginner

4

Introduction to NIM Microservices

How you actually serve a model to a Flask app via an inference endpoint.

~1 hr · Beginner

5

Building RAG Agents with LLMs — the centrepiece

Highest-value free course for you. LangChain · FAISS · LangServe · FastAPI — maps almost 1:1 onto how you build.

~8 hr · Intermediate · certificate

6

Agentic AI Explained

Extends RAG into tool-using agents.

~1–2 hr · Beginner

7

Then breadth, as needed

Even Easier Intro to CUDA · Accelerate Data Science (Zero Code) · AI in the Data Center · Digital Fingerprinting.

optional depth

Generative AI · LLMs · RAG

6 free

Deep Learning & fundamentals

4 free · 1 paid sibling

Accelerated computing & data science

2 free · 2 paid siblings

An Even Easier Introduction to CUDA

The free CUDA on-ramp — your first GPU-accelerated program.

FREE~1 hrBeginner

Accelerate Data Science — Zero Code Changes

Speed up pandas/DataFrames with RAPIDS, no rewrites. The free DS taster.

FREE~1 hrBeginner

Infrastructure · security · robotics · Omniverse

7+ free

Breadth for when you want it — deployment literacy, cybersecurity, edge robotics, and 3D/simulation.