Module 1: Advanced AI Workload Architectures

• The private cloud enterprise—the bare metal, VMs, and containers
• The private cloud enterprise—from DevOps to analytics
• GreenLake for Private Cloud Enterprise
• Multi-node training and distributed inference
• Hybrid AI deployment models (on-prem + cloud)
• Advanced workload classification (fine-tuning, RAG+, multi-modal AI)
• Designing AI pipelines for enterprise scale

Module 2: Advanced Data Engineering for AI

• Feature engineering at enterprise scale
• Data versioning and lineage
• Handling images, audio, and video
• Feature stores and real-time data ingestion

Module 3: Multi-Modal AI Applications

• Fundamentals of multi-modal architectures
• Multi-modal transformers and embeddings
• Use cases: document intelligence, vision-language models
• Integration with enterprise data sources

Module 4: Fine-Tuning and Transfer Learning

• Review of pre-trained model adaptation
• LoRA, QLoRA, and adapters
• Fine-tuning vs prompt-tuning vs distillation
• Case studies for customizing enterprise models

Module 5: Retrieval-Augmented Generation at Scale

• Hierarchical retrieval and multi-vector indexing
• Large-scale vector databases
• Query expansion and relevance tuning
• Hybrid retrieval (dense + sparse)

Module 6: Orchestration and AIOps with HPE OpsRamp

• Multi-agent collaboration patterns
• Orchestration frameworks (LangGraph, AutoGen)
• Memory, planning, and tool use in agents
• Enterprise scenarios: incident response, alert triage, compliance enforcement, resource optimization with HPE OpsRamp
• Proactive capacity planning and performance forecasting using HPE OpsRamp
• Automated SLA/SLO monitoring and service health scoring
• AIOps-driven resource scaling and optimization recommendations

Module 7: Advanced Automation and MLOps

• CI/CD for AI pipelines
• Model registries and governance
• Continuous validation and drift detection
• Automation frameworks for lifecycle management
• Governance and access controls with HPE Morpheus Software role-based orchestration
• Cost, quota, and lifecycle automation through HPE Morpheus Software policies
• HPE Morpheus Software analytics for hybrid-cloud observability
• Blueprint-based deployment of consistent ML environments across clouds and on-prem
• API-driven automation for provisioning, deprovisioning, and updating model-serving environments

Module 8: NVIDIA NIM Foundations and Enterprise Model Deployment

• NVIDIA inference stack: Triton, TensorRT-LLM, NeMo, CUDA
• NIM microservices: LLM, RAG, multimodal, embeddings, ASR/TTS
• Deploying NIM containers via Docker or Helm charts
• GPU optimization and performance tuning for NIM services
• Enterprise integration using REST and gRPC endpoints
• Monitoring and observability: DCGM, Prometheus, NVML

Module 9: Building Enterprise RAG Pipelines with NVIDIA NIM

• NVIDIA RAG reference architecture
• NIM embedding and reranker microservices
• Hybrid retrieval design: dense + sparse + vector fusion
• GPU-accelerated vector search using RAFT and cuVS
• Integration with Triton Inference Server
• Large-scale RAG orchestration patterns for enterprise workloads

Module 10: NVIDIA NeMo for Fine-Tuning and Model Customization

• NeMo architecture: collections, configs, Lightning integration
• Distributed training with Megatron-LM (TP/PP/DP)
• Dataset preparation, tokenization, and pre-processing pipelines
• Parameter-efficient fine-tuning workflows
• Model checkpointing, evaluation, and export

Module 11: NeMo Guardrails for Enterprise AI Governance

• Overview of NeMo Guardrails and RailSpec syntax
• Building rule sets for enterprise applications
• Grounding, safety, and privacy rule enforcement
• Logging, evaluation, and continuous guardrail optimization
• Integration with LLM apps, RAG pipelines, and agents