What is the NVIDIA certification path?
Every figure on this page is transcribed from NVIDIA’s own certification pages and was last checked on . Prices and policies are set by NVIDIA and can change — verify on nvidia.com before booking. passgenai is an independent study resource and is not affiliated with, endorsed by, or sponsored by NVIDIA Corporation.
The four ladders
Each track has its own associate and professional tier and they run independently — nothing in one track gates anything in another. Associate first within your track, professional once you have shipped the work.
Generative AI
Building with large language models, multimodal models and agents — prompting, retrieval, fine-tuning, evaluation and serving. This is the track for application and ML engineers.
- NCA-GENLAssociate · $125 · 1 hour
NVIDIA Certified Associate: Generative AI LLMs
NVIDIA assumes: A basic understanding of generative AI and large language models
- NCA-GENMAssociate · $125 · 1 hour
NVIDIA Certified Associate: Generative AI Multimodal
NVIDIA assumes: A basic understanding of generative AI
- NCP-GENLProfessional · $200 · 2 hours
NVIDIA Certified Professional: Generative AI LLMs
NVIDIA assumes: 2–3 years of practical AI/ML experience with large language models, including transformer architectures, prompt engineering, distributed parallelism and parameter-efficient fine-tuning
- NCP-AAIProfessional · $200 · 2 hours
NVIDIA Certified Professional: Agentic AI
NVIDIA assumes: 1–2 years in AI/ML roles with hands-on work on production agentic AI projects — agent architecture, orchestration, multi-agent frameworks, evaluation, observability and guardrails
AI Infrastructure
Specifying, building and running the GPU estate underneath the models — servers, networking, racks, power, cooling, orchestration and monitoring. This is the track for infrastructure, platform and data-centre engineers.
- NCA-AIIOAssociate · $125 · 1 hour
NVIDIA-Certified Associate: AI Infrastructure and Operations
NVIDIA assumes: A basic understanding of data center infrastructure
- NCP-AIIProfessional · $400 · 2 hours
NVIDIA-Certified Professional: AI Infrastructure
- NCP-AINProfessional · $400 · 2 hours
NVIDIA-Certified Professional: AI Networking
- NCP-ARIProfessional · $400 · 2 hours
NVIDIA-Certified Professional: AI Rack and Interconnect
- NCP-AIOProfessional · $500 · 2 hours
NVIDIA-Certified Professional: AI Operations
Data Science
GPU-accelerated data science: moving dataframe, machine-learning and graph workloads onto the GPU with the RAPIDS stack instead of scaling out on CPUs.
- NCA-ADSAssociate · $125 · 1 hour
NVIDIA-Certified Associate: Accelerated Data Science
- NCP-ADSProfessional · $200 · 2 hours
NVIDIA-Certified Professional: Accelerated Data Science
Simulation and Physical AI
Describing and composing 3D worlds with OpenUSD — the format underneath digital twins, simulation and robotics pipelines.
- NCP-OUSDProfessional · $200 · 2 hours
NVIDIA-Certified Professional: OpenUSD Development
Note the gap: this track has no associate exam, so the professional OpenUSD exam is the only entry point.
Choosing your first exam in one question
What do you spend most of your week doing?
- Writing code that calls models — prompts, RAG, agents, evaluation
- NCA-GENL
The associate generative-AI exam. It is the default first exam in the programme and the one whose vocabulary the others assume.
- Working with images, audio or video as much as text
- NCA-GENM
Same tier and same fee as NCA-GENL, but the domains are spread across modalities rather than concentrated on language.
- Specifying, buying, racking or running GPU capacity
- NCA-AIIO
The associate infrastructure exam. Forty per cent of it is AI infrastructure proper — sizing, power, cooling, networking — which is the part your job already argues about.
- Data science at a scale where CPUs have become the bottleneck
- NCA-ADS
The accelerated data-science associate exam, built around the RAPIDS stack and the question of when moving a workload to a GPU pays for itself.
- Already operating LLM systems in production, for years
- NCP-GENL or NCP-AAI
Skip the associate tier. NCP-GENL if your work is model optimisation, serving and fine-tuning; NCP-AAI if it is agents, orchestration and guardrails.
- 3D, digital twins, robotics simulation
- NCP-OUSD
The OpenUSD professional exam, and the only exam in its track — there is no associate rung to warm up on.
Associate or professional: how to tell which you are ready for
The tiers differ in more than difficulty, and the difference is not one you can bridge with extra revision.
Associate exams test whether you know the landscape. One hour, $125, and the questions ask what a technique is for, what it trades away and when it is the wrong tool. A capable beginner with a few weeks of structured study passes one. The prerequisites NVIDIA states are deliberately soft — “a basic understanding of” — and they mean it.
Professional exams test whether you have done the job. Two hours, and NVIDIA states one to three years of hands-on experience depending on the exam. The questions are scenario-shaped: a system, a constraint, a symptom, and four plausible responses of which one is what an engineer who has been on call would do. That is not knowledge you can revise into place in a fortnight, which is the honest reason to sit the associate exam first if you are unsure.
A useful self-test: read the domain list for the professional exam you are considering. If more than two domains describe work you have never personally done, you are looking at the wrong tier — and a failed attempt costs the fee again plus a 14-day wait.
What NVIDIA does not have
Worth naming, because other vendors do have these and people arrive expecting them.
- No gated prerequisites. No exam requires you to hold another one first. You could sit the most expensive professional exam as your first ever NVIDIA exam — inadvisable, but permitted.
- No expert or architect tier. The programme stops at professional. There is no third rung above it, so “NVIDIA Certified Expert” is not a thing that exists.
- No specialty add-ons. Unlike the cloud vendors, NVIDIA does not layer specialty certifications on top of a core one. Each exam stands alone.
- No continuing-education renewal. Credentials last two years and you recertify by retaking the exam, not by accumulating credits. That makes collecting certifications expensive to sustain and argues for holding one or two that match your work.
Related questions
What is the NVIDIA certification path?
There is no mandatory sequence — every NVIDIA exam is open to anyone, and no certification is a prerequisite for another. What exists is a sensible order, and it is decided by your track rather than by the programme. Pick the track that matches your work: Generative AI if you build with models, AI Infrastructure if you run the GPUs underneath them, Data Science if your problem is dataframe and machine-learning workloads at scale, Simulation if you work in OpenUSD and digital twins. Then sit that track's associate exam first ($125, one hour) and its professional exam once you have real delivery experience — the professional papers assume one to three years of hands-on work and are written for someone who has operated the systems they describe, so taking one early is an expensive way to discover that. The one genuine dependency is conceptual: NCP-AAI builds on LLM fundamentals, so NCA-GENL or equivalent experience should come first.
Which NVIDIA certification should you take first?
For most people, NCA-GENL — the associate generative-AI exam. It is the cheapest tier at $125, assumes only a basic understanding of LLMs, and its syllabus is the vocabulary the rest of the programme takes for granted. Two exceptions worth naming: if your job is infrastructure rather than application work, start with NCA-AIIO instead, and if you work mainly with images, audio or video, NCA-GENM is the closer match. Do not start with a professional exam.
How many NVIDIA certifications are there?
There are 12 NVIDIA certifications, across four tracks: four in Generative AI (NCA-GENL, NCA-GENM, NCP-GENL, NCP-AAI), five in AI Infrastructure (NCA-AIIO, NCP-AII, NCP-AIO, NCP-AIN, NCP-ARI), two in Data Science (NCA-ADS, NCP-ADS) and one in Simulation and Physical AI (NCP-OUSD). Four are associate-level and eight professional.
Is there an NVIDIA certification for beginners?
The associate tier is the beginner tier, and its prerequisites are deliberately light — NCA-GENL asks for a basic understanding of generative AI and LLMs, NCA-AIIO for a basic understanding of data-centre infrastructure. Neither expects professional experience. They are not trivial exams: they cover a broad syllabus in an hour and reward judgement over recall. But a motivated beginner with a few weeks of structured study can pass one, which is not true of any professional exam in the programme.