NCA-AIIO AI Infrastructure Operations Associate Study Plan¶

6-Week Study Schedule¶

Week 1: GPU Fundamentals and Monitoring¶

• Study GPU architecture basics (SMs, Tensor Cores, HBM)
• Learn nvidia-smi commands and output interpretation
• Practice DCGM commands (diag, dmon, health)
• Understand ECC errors (SBE vs DBE)
• Reference: DCGM Docs

Week 2: Container and Runtime Management¶

• Install and configure NVIDIA Container Toolkit
• Practice Docker --gpus flags
• Learn NGC container registry usage
• Study container best practices for GPU
• Reference: Container Toolkit Docs

Week 3: Kubernetes GPU Operations¶

• Study GPU Operator components
• Practice GPU scheduling in pod specs
• Learn MIG basics and resource types
• Understand GPU Feature Discovery labels
• Reference: GPU Operator Docs

Week 4: Infrastructure and Troubleshooting¶

• Study GPU driver management
• Learn common GPU issues and resolution
• Practice log analysis (dmesg, Xid errors)
• Understand maintenance procedures
• Lab: Practice troubleshooting scenarios

Week 5: Review and Practice¶

• Work through scenario-based questions
• Review all domains
• Create flashcards for commands and error codes
• Focus on weak areas

Week 6: Final Preparation¶

• Timed practice session
• Review fact sheet
• Exam logistics preparation

Self-Assessment¶

• Can I interpret nvidia-smi output?
• Do I know DCGM diagnostic levels and commands?
• Can I run Docker containers with GPU access?
• Do I know GPU Operator components?
• Can I troubleshoot common GPU issues?