# ML in Production Practice

## Docs

- [Containerization & Infrastructure](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/concepts/containerization.md): Understanding containerization fundamentals and Kubernetes for ML deployments
- [Data Management](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/concepts/data-management.md): Storage, versioning, labeling, and efficient processing of ML datasets
- [Model Serving](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/concepts/model-serving.md): Deploying ML models as production APIs with FastAPI, Triton, vLLM, and KServe
- [Monitoring & Observability](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/concepts/monitoring.md): Track ML system health, detect drift, and debug production issues
- [Model Optimization](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/concepts/optimization.md): Quantization, scaling strategies, and performance tuning for production ML systems
- [Pipeline Orchestration](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/concepts/pipeline-orchestration.md): Automating ML workflows with Airflow, Kubeflow, and Dagster
- [Production Patterns](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/concepts/production-patterns.md): Buy vs build decisions, cloud platforms, and real-world ML system design
- [Training Workflows](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/concepts/training-workflows.md): Experiment tracking, configuration management, and reproducible ML training
- [Introduction](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/introduction.md): Practical exercises and reference implementations for Machine Learning in Production
- [CI/CD with GitHub Actions](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-1/cicd.md): Automate building, testing, and deploying ML containers with GitHub Actions
- [Docker Containerization](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-1/docker.md): Build, run, and share Docker images for ML applications
- [Kubernetes Deployment](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-1/kubernetes.md): Deploy and orchestrate containerized ML workloads with Kubernetes
- [Module 1: Infrastructure Overview](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-1/overview.md): Learn containerization, Kubernetes orchestration, and CI/CD pipelines for ML systems
- [Practice Exercise](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-1/practice.md): Hands-on tasks to master containerization, Kubernetes, and CI/CD
- [Serverless Alternatives](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-1/serverless.md): Deploy ML applications without managing infrastructure using Modal and Railway
- [Data Formats](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-2/formats.md): Compare storage formats and optimize data loading performance
- [Data Labeling](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-2/labeling.md): Deploy Argilla for annotation and create synthetic datasets
- [Module 2: Data Management](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-2/overview.md): Master data storage, processing, and management for production ML systems
- [Practice Tasks](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-2/practice.md): Hands-on exercises for Module 2: Data Management
- [Data Storage](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-2/storage.md): Deploy MinIO, implement S3 clients, and manage datasets with DVC
- [Streaming Datasets](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-2/streaming.md): Create and consume streaming datasets for efficient ML training
- [Vector Databases](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-2/vector-databases.md): Build RAG applications with LanceDB and semantic search
- [Classic Training (BERT)](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-3/classic-training.md): Train BERT-based text classification models with HuggingFace Transformers
- [Experiment Tracking](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-3/experiment-tracking.md): Track, compare, and manage ML experiments with modern MLOps tools
- [LLM Training (Phi-3)](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-3/llm-training.md): Fine-tune large language models with LoRA and parameter-efficient techniques
- [Model Cards](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-3/model-cards.md): Document ML models with standardized model cards for transparency and accountability
- [Module 3: Training Workflows](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-3/overview.md): Learn to structure ML training projects, track experiments, and manage configurations
- [Practice Exercises](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-3/practice.md): Hands-on homework assignments for Module 3: Training Workflows
- [Project Structure](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-3/project-structure.md): Organize ML training projects with proper Python packaging and structure
- [Apache Airflow Pipelines](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-4/airflow.md): Build ML pipelines with Airflow and KubernetesPodOperator
- [Dagster Pipelines](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-4/dagster.md): Build asset-centric ML pipelines with data quality checks
- [Kubeflow Pipelines](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-4/kubeflow.md): Build Kubernetes-native ML pipelines with artifact tracking
- [Pipeline Orchestration Overview](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-4/overview.md): Learn to orchestrate ML pipelines using Airflow, Kubeflow, and Dagster
- [Practice Exercises](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-4/practice.md): Hands-on exercises for building ML pipelines with Airflow, Kubeflow, and Dagster
- [FastAPI Model Serving](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-5/fastapi.md): Build production-ready REST APIs with FastAPI and Pydantic
- [KServe Inference Serving](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-5/kserve.md): Deploy cloud-native model serving with KServe on Kubernetes
- [Module 5: Model Serving](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-5/overview.md): Deploy production-ready APIs, UIs, and inference servers for ML models
- [Module 5 Practice](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-5/practice.md): Hands-on tasks for model serving with APIs, UIs, and inference servers
- [Streamlit UI Serving](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-5/streamlit.md): Build interactive web interfaces for ML models with Streamlit
- [Triton Inference Server](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-5/triton.md): Deploy high-performance model serving with NVIDIA Triton
- [vLLM Serving for LLMs](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-5/vllm.md): Serve large language models with vLLM and dynamic LoRA adapter loading
- [Async Inference](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-6/async-inference.md): Implement queue-based inference patterns for better scalability
- [Autoscaling](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-6/autoscaling.md): Configure HPA and KNative autoscaling for ML workloads
- [Load Testing](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-6/load-testing.md): Benchmark your ML APIs with Locust, k6, and Vegeta
- [Module 6: Optimization](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-6/overview.md): Learn benchmarking, autoscaling, and model optimization techniques for production ML systems
- [Practice Exercises](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-6/practice.md): Hands-on practice for Module 6: Optimization
- [Model Quantization](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-6/quantization.md): Optimize inference performance with quantization techniques
- [Data Monitoring](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-7/data-monitoring.md): Monitor ML models with Evidently and Seldon for drift detection and outlier analysis
- [Grafana](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-7/grafana.md): Set up Prometheus and Grafana for Kubernetes metrics and custom dashboards
- [LLM Observability](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-7/observability.md): Instrument LLM applications with OpenTelemetry and specialized observability tools
- [Module 7: Monitoring](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-7/overview.md): Set up observability and monitoring for ML applications in production
- [Practice](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-7/practice.md): Hands-on homework assignments for monitoring and observability
- [SigNoz](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-7/signoz.md): Deploy and configure SigNoz for distributed tracing and application monitoring
- [AWS SageMaker Multi-Model Endpoints](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-8/aws-sagemaker.md): Deploy multiple models behind a single endpoint for cost-efficient inference
- [Buy vs Build Decision Framework](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-8/buy-vs-build.md): Structured approach to evaluating whether to build custom ML infrastructure or adopt managed platforms
- [Module 8: Cloud Platforms](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-8/overview.md): Explore buy vs build decisions and deploy multi-model endpoints on cloud platforms
- [Practice: Cloud Platforms](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/modules/module-8/practice.md): Compare managed ML platforms and implement multi-model deployments
- [Quickstart](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/quickstart.md): Get started with ML in Production Practice in under 10 minutes
- [Setup](https://mintlify.wiki/kyryl-opens-ml/ml-in-production-practice/setup.md): Detailed environment setup guide for ML in Production Practice