03 - MLOps & Infrastructure Index

MLOps & Infrastructure

---

Model Lifecycle Management

Version Control

• ML Version Control Fundamentals
• Model Versioning
• Dataset Versioning
• Experiment Tracking
• DVC (Data Version Control)

Experiment Management

• Experiment Tracking Best Practices
• MLflow
• Weights & Biases
• Neptune.ai
• Reproducibility in ML
• Hyperparameter Tracking

Model Registry

• Model Registry Concepts
• Model Metadata Management
• Model Lineage Tracking
• Model Governance

Deployment Pipelines

• ML Deployment Patterns
• CI CD for ML
• Shadow Deployment
• Rolling Deployment

Monitoring & Observability

• ML Monitoring Fundamentals
• Model Performance Monitoring
• Data Drift Detection
• Concept Drift
• Feature Drift
• Model Staleness
• Logging for ML Systems
• Alerting and Incident Response

---

Infrastructure & Scaling

Compute Resources

• GPU Fundamentals for ML
• TPU Overview
• GPU Utilization Optimization
• Multi-GPU Training
• GPU Memory Management
• Mixed Precision Training

Distributed Training

• Distributed Training Overview
• Data Parallelism
• Model Parallelism
• Pipeline Parallelism
• Distributed Data Parallel (DDP)
• Fully Sharded Data Parallel (FSDP)
• Horovod
• DeepSpeed

Model Optimization

• Model Compression Techniques
• Quantization
• Pruning
• Knowledge Distillation
• ONNX Runtime
• TensorRT
• Neural Architecture Search

Edge Deployment

• Edge ML Fundamentals
• TensorFlow Lite
• ONNX
• CoreML
• Model Optimization for Edge
• On-Device Inference

Cloud Platforms

• Cloud ML Platforms Overview
• AWS SageMaker
• Azure Machine Learning
• Google Cloud Vertex AI
• Databricks ML
• Serverless ML Deployment

Container & Orchestration

• Docker for ML
• Kubernetes for ML
• Kubeflow
• MLOps Pipeline Architecture
• Model Serving with Kubernetes

Storage & Databases

• ML Storage Strategies
• Feature Stores
• Object Storage for ML
• Model Artifact Storage

---

📊 Progress Tracking

TABLE
  status as "Status",
  difficulty as "Difficulty",
  last_modified as "Last Updated"
FROM "01 - ML & AI Concepts/03 - MLOps & Infrastructure"
WHERE contains(tags, "concept")
SORT file.name ASC

---

🎓 Learning Path

Recommended Order:

1. Start with Model Lifecycle basics (Version Control, Experiment Tracking)
2. Learn Deployment Patterns
3. Study Monitoring & Observability
4. Understand Compute Resources
5. Master Model Optimization
6. Explore Cloud Platforms
7. Advanced: Distributed Training and Edge Deployment

---

Back to: ML & AI Index