Durga Analytics training

NVIDIA certification training

Turn a NVIDIA credential into capability your team can actually deploy, and a career step that stands up at interview.

We run instructor-led NVIDIA programmes for individuals and for whole teams: a platform mastery track organised around running the platform, and certification programmes focused on one named credential and its published exam objectives. Both are taught by practitioners who have run these platforms in production. The full NVIDIA credential reference sits at the bottom of this page.

11 programmes we teach · 11 credentials referenced

Start here

The programme most teams should take first

Our training is not organised around an exam date. We teach the platform properly and the certifications follow, because the syllabus is a superset of what they test. If you are not sure which credential to chase, start with the mastery track.

Start here · no exam required · DA-200 to DA-400

NVIDIA AI Infrastructure Mastery

GPU estates fail on networking, scheduling and thermals long before they fail on model code. A working command of GPU infrastructure and the NVIDIA software stack, covering the ground the NCA and NCP credentials examine.

How we teach

What NVIDIA training changes

You build, we review

Every module runs a guided lab and then an independent challenge where the requirement changes or a fault is introduced. Your implementation is reviewed with an instructor, who questions the decisions rather than marking a checklist.

How it is taught

Sessions are live, questions are answered in the session, and the teaching is built around the failure modes that actually bite in production.

Assessed on what you can do

The programme closes with a practical task and a conversation about the trade-offs you chose. That is the evidence of capability, which is a different thing from an exam score.

Independent by design

We resell nobody's curriculum and carry no partner quota, so the advice can be that a credential is not worth your quarter. That is worth more to you than another seat sold.

Certification programmes

NVIDIA certification programmes

Organised by the role they serve and placed on the Durga ladder. Each one is instructor-led, as a scheduled cohort or delivered privately to your team. Register your interest and we will come back with dates, scope and pricing.

Data Scientist

2 programmes

Generative AI Engineer

3 programmes

Agentic AI Engineer

1 programme

AI Infrastructure Engineer

2 programmes

AI Operations Engineer

1 programme

Network Engineer

1 programme

3D/Simulation Developer

1 programme
Reference

Every NVIDIA credential

The full catalogue as NVIDIA publishes it, so you can see where our programmes sit. Click any exam code to open the official page. Request training on any credential here, including the ones we have not yet built a full programme around.

AI Infrastructure

CertificationDurga levelExamTargetsPrice (USD)ValidTraining
Certified Associate AI Infrastructure and OperationsOne hour.DA-200 PractitionerNCA-AIIONot published by NVIDIA125 USD2 yearsRegister Training Team training Programme
Certified Professional AI InfrastructureTwo hours.DA-300 ProfessionalNCP-AIINot published by NVIDIA400 USD2 yearsRegister Training Team training Programme
Certified Professional AI OperationsTwo hours.DA-400 SpecialistNCP-AIONot published by NVIDIA500 USD2 yearsRegister Training Team training Programme
Certified Professional AI NetworkingTwo hours.DA-400 SpecialistNCP-AINNot published by NVIDIA400 USD2 yearsRegister Training Team training Programme

Data Science

CertificationDurga levelExamTargetsPrice (USD)ValidTraining
Certified Associate Accelerated Data ScienceOne hour.DA-200 PractitionerNCA-ADSNot published by NVIDIA125 USD2 yearsRegister Training Team training Programme
Certified Professional Accelerated Data ScienceTwo hours.DA-300 ProfessionalNCP-ADSNot published by NVIDIA200 USD2 yearsRegister Training Team training Programme

Generative AI

CertificationDurga levelExamTargetsPrice (USD)ValidTraining
Certified Associate Generative AI LLMOne hour.DA-200 PractitionerNCA-GENLNot published by NVIDIA125 USD2 yearsRegister Training Team training Programme
Certified Associate Generative AI MultimodalOne hour.DA-200 PractitionerNCA-GENMNot published by NVIDIA125 USD2 yearsRegister Training Team training Programme
Certified Professional Generative AI LLMsTwo hours.DA-300 ProfessionalNCP-GENLNot published by NVIDIA200 USD2 yearsRegister Training Team training Programme
Certified Professional Agentic AITwo hours.DA-300 ProfessionalNCP-AAINot published by NVIDIA200 USD2 yearsRegister Training Team training Programme

OpenUSD

CertificationDurga levelExamTargetsPrice (USD)ValidTraining
Certified Professional OpenUSD DevelopmentTwo hours.DA-400 SpecialistNCP-OUSDNot published by NVIDIA200 USD2 yearsRegister Training Team training Programme
Reference

How NVIDIA structures its programme

NVIDIA runs a two-level programme organised by technology domain rather than by job role. Associate exams run one hour with roughly 40 to 60 questions. Professional exams run two hours and some expect hands-on experience with NVIDIA technologies. Exams are grouped into published areas: AI Infrastructure, Data Science, Generative AI and OpenUSD, plus a Simulation and Physical AI heading that currently lists no certifications. Exams are pass or fail with no score reported, delivered in person or remotely through NVIDIA Authorized Testing Partners with proctoring and identity checks, and digital credentials are issued through Credly within 24 hours.

Renewal, validity and cost

NVIDIA certifications are valid for two years, after which you must retake the exam to be recertified. There is no continuing-education alternative. Retake rules are a 14-day wait between attempts and a maximum of five attempts per exam in any 12 months.

On price

NVIDIA publishes USD prices on the catalogue page. It notes that prices can change, that taxes may apply, and that educational and group discounts are available.

Reference

Before you book an exam

  • NVIDIA does not publish a target job role per exam. It advises choosing by job role, expertise and career goals, which is why the role column reads as not published.
  • The eleven credentials above are NVIDIA's complete published catalogue as shown on its certification page.