Course Overview
- Course Title: NCA‑AIIO SoAI‑Certified Associate: AI Infrastructure & Operations
- Instructor: School of AI (AI Academy)
- Target Audience:
- IT professionals managing GPU-accelerated infrastructure
- System administrators and DevOps engineers deploying AI workloads
- AI/ML engineers transitioning to infrastructure roles
- Cloud architects optimizing AI pipelines
- Students/enthusiasts preparing for NVIDIA NCA-AIIO certification
- Prerequisites:
- Basic knowledge of IT systems (servers, storage, networking, or Linux)
- Familiarity with cloud platforms (AWS/Azure/GCP) helpful but not required
- No prior GPU/AI experience needed
Curriculum Highlights
-
Key Topics Covered:
- GPU Acceleration for AI: Tensor Cores, Streaming Multiprocessors (SMs), NVLink, MIG (Multi-Instance GPU)
- AI Lifecycle Management: Model development → training → deployment → monitoring → scaling
- Tools & Platforms:
- NVIDIA NGC, Triton Inference Server, DCGM (Data Center GPU Manager)
- Kubeflow, MLflow, Helm Charts, TensorRT
- GPU-Accelerated Infrastructure:
- GPUDirect, RDMA, InfiniBand vs. Ethernet
- vGPUs, multi-tenant deployments, BlueField DPUs, DOCA SDK
- Storage & Networking: GPU-optimized storage, NVSwitch, GPU clustering
- Exam Preparation: Full-length mock test (50 questions), flashcards, readiness checklist
-
Key Skills Learned:
- Deploy and monitor AI workloads using DCGM and NGC
- Configure GPU clusters for scalable AI training/inference
- Optimize MLOps pipelines with Kubeflow/MLflow
- Implement GPUDirect RDMA and vGPU for multi-tenancy
- Compare InfiniBand vs. Ethernet for AI networking
- Prepare for NVIDIA NCA-AIIO certification exam
Course Format
- Duration: 2.5 hours on-demand video
- Format: Self-paced online course (pre-recorded lectures + hands-on labs)
- Resources:
- 6 articles
- 7 downloadable resources (labs, checklists, flashcards)
- Mobile/TV access
- Certificate of completion
- Hands-on labs (Google Colab/NGC-based simulations)
Additional Details
- Exam Alignment: 100% mapped to NVIDIA’s official NCA-AIIO blueprint
- GPU Models Covered: A100, H100, L40s, B200
- Hands-on Tools: NGC Catalog, Triton Inference Server, DCGM, Helm Charts, DOCA SDK
- Career Focus: AI infrastructure engineering, GPU cluster administration, MLOps operations