Gerçek Zamanlı Yapay Zeka Proje Yönetimi
MLOps ve Model Deployment

MLOps Temelleri ve Model Lifecycle

Hafta 1 konuları

• • MLOps prensipleri ve DevOps'tan farkları
• • Model development lifecycle ve stage'leri
• • Technical debt ve model governance
• • Reproducibility ve experiment tracking temelleri

Containerization ve Docker

Hafta 2 konuları

• • Docker fundamentals ve Dockerfile best practices
• • Model serving container'ları oluşturma
• • Multi-stage builds ve image optimization
• • Docker registry ve versiyonlama stratejileri

Kubernetes ve Orchestration

Hafta 3 konuları

• • Kubernetes architecture ve core components
• • Deployment, Service, Ingress configuration
• • Horizontal Pod Autoscaling (HPA) ve load balancing
• • StatefulSets ve persistent storage for models

CI/CD Pipeline'ları

Hafta 4 konuları

• • GitHub Actions ve GitLab CI/CD for ML
• • Jenkins pipeline ve automated testing
• • Model validation ve smoke testing
• • Automated deployment strategies (blue-green, canary)

Cloud ML Platformları - Azure

Hafta 5 konuları

• • Azure Machine Learning workspace setup
• • Azure ML pipelines ve compute clusters
• • Model registry ve endpoint deployment
• • Azure DevOps integration

AWS SageMaker ve Google Cloud AI

Hafta 6 konuları

• • SageMaker training jobs ve endpoints
• • SageMaker pipelines ve model monitor
• • Google Cloud AI Platform ve Vertex AI
• • Multi-cloud deployment stratejileri

Model Serving ve API Development

Hafta 7 konuları

• • FastAPI ve Flask for model APIs
• • gRPC ve protocol buffers
• • Batch inference ve streaming predictions
• • API gateway ve rate limiting

Monitoring ve Observability

Hafta 8 konuları

• • Prometheus metrics collection
• • Grafana dashboard design
• • Data drift ve concept drift detection
• • Model performance degradation alerts

MLflow ve Kubeflow

Hafta 9 konuları

• • MLflow tracking server setup
• • Model registry ve versioning
• • Kubeflow pipelines ve components
• • Distributed training orchestration

Edge Deployment ve Optimization

Hafta 10 konuları

• • Model compression ve quantization
• • ONNX ve TensorFlow Lite conversion
• • Edge device deployment strategies
• • Real-time inference optimization

Uygulama Modülü 1
End-to-End Computer Vision Pipeline

Hafta 11 konuları

• • Image classification model training ve optimization
• • Multi-stage Docker container build
• • Kubernetes deployment with autoscaling
• • Monitoring dashboard ve alerting setup

Uygulama Modülü 2
NLP Model Production System

Hafta 12 konuları

• • Text processing pipeline development
• • A/B testing infrastructure setup
• • Rolling update ve rollback strategies
• • Performance monitoring ve cost optimization