Platform mastery · no exam required

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. This is platform mastery, not certification preparation: it is organised around implementing, operating and improving the platform. Certification is an optional outcome, and the section below sets out exactly which credentials this track prepares you for and which ones need their own programme.

DA-200 to DA-400 · 7 modules · cohort or private team delivery

The programme

What it covers

Built around running the platform, not around an exam blueprint. Each module ends in applied work on a real estate, and the last one is a capstone you can show your own leadership.

  1. GPU architecture and selection

    Streaming multiprocessors, memory bandwidth and capacity, NVLink, and matching silicon to training versus inference workloads.

  2. Cluster and network design

    Multi-GPU and multi-node topology, InfiniBand and RoCE, collective communication, and the bottlenecks that waste the most expensive hardware you own.

  3. The software stack

    CUDA and drivers, containers and NGC, the GPU Operator, and Kubernetes scheduling for GPU workloads.

  4. Operations and multi-tenancy

    MIG partitioning, job scheduling and queueing, utilisation telemetry, health monitoring and capacity planning.

  5. Inference and serving

    Triton Inference Server, TensorRT optimisation, batching strategy, and the latency and throughput trade-off.

  6. Power, cooling and economics

    Power and thermal envelopes, rack density, and the operating cost of an owned GPU estate against cloud alternatives.

  7. Capstone

    Size, stand up and instrument a GPU cluster for a stated workload, then defend the design against a cloud comparison.

How it is delivered

Two ways to take it

Both tiers are live, instructor-led delivery. Schedule and depth are agreed before the programme starts rather than fixed in advance.
FeatureCohortEnterprise
FormatScheduled live sessions with a mixed peer groupThe same programme delivered privately to your team
Live sessionsInstructor-led, scheduledInstructor-led, scheduled around your delivery calendar
LabsShared lab scenariosRun against your own estate where access allows
TailoringFixed syllabusSequenced to your platforms, workloads and governance
Exam focusBlueprint coverage plus question-format drillingBlueprint coverage, plus the gaps found in your team's baseline
Group sizeIndividual seatsWhole team, one cohort
Best forOne person preparing around a jobA team standardising on the same platform
Where it leads

Which certifications this track supports

A seven-module platform track cannot make anyone advanced across every role the vendor certifies, so we say which is which. You do not have to sit any of them.

Covered by this track

The platform ground these credentials test is covered here. Add focused exam drilling for the format and you are ready to sit.

Needs its own programme

This track gives you the platform foundation these build on, but each is a separate professional specialism and needs its own preparation. We run a programme for each.

Take the track

Register and we will come back with cohort dates, scope and pricing. For a team, tell us how many people and which workloads you run and we will scope a private delivery on your own estate.

Independent training. Durga Analytics is not a vendor partner and does not award credentials. Verify exam details officially. Credential detail reviewed 13 October 2026.