# Machine Learning Design Patterns — designpattern.fyi > 30 patterns across the ML lifecycle - from data representation to responsible AI - **Section**: Machine Learning Design Patterns - **Canonical URL**: https://designpattern.fyi/machine_learning/ --- ## Patterns in this Category ### [Batch Serving](https://designpattern.fyi/machine_learning/batch-serving/) - **Summary**: Precomputes predictions asynchronously using distributed data processing - **Subcategory**: Serving and Operational Resilience ### [Bridged Schema](https://designpattern.fyi/machine_learning/bridged-schema/) - **Summary**: Forward-fits old-format data into new schemas to preserve historical training data - **Subcategory**: Training Process and Optimization ### [Cascade](https://designpattern.fyi/machine_learning/cascade/) - **Summary**: Decomposes difficult problems into a sequence of smaller, more homogeneous ML problems - **Subcategory**: Problem Framing and Model Structure ### [Checkpoints](https://designpattern.fyi/machine_learning/checkpoints/) - **Summary**: Periodic saving of model state during training for fault tolerance and early stopping - **Subcategory**: Training Process and Optimization ### [Continuous Model Evaluation](https://designpattern.fyi/machine_learning/continuous-model-evaluation/) - **Summary**: Ongoing pipeline to detect model performance degradation over time - **Subcategory**: Serving and Operational Resilience ### [Distribution Strategy](https://designpattern.fyi/machine_learning/distribution-strategy/) - **Summary**: Scales training across multiple processors or machines through parallelism - **Subcategory**: Training Process and Optimization ### [Embedding](https://designpattern.fyi/machine_learning/embedding/) - **Summary**: Dense, lower-dimensional vector representations that capture latent similarity - **Subcategory**: Data and Feature Representation ### [Ensemble](https://designpattern.fyi/machine_learning/ensemble/) - **Summary**: Combines predictions from multiple models to improve accuracy and robustness - **Subcategory**: Problem Framing and Model Structure ### [Explainable Predictions](https://designpattern.fyi/machine_learning/explainable-predictions/) - **Summary**: Provides human-interpretable explanations for individual model predictions - **Subcategory**: Trust, Evaluation and Responsible AI ### [Fairness Lens](https://designpattern.fyi/machine_learning/fairness-lens/) - **Summary**: Evaluates and designs systems for equitable treatment across relevant subgroups - **Subcategory**: Trust, Evaluation and Responsible AI ### [Feature Cross](https://designpattern.fyi/machine_learning/feature-cross/) - **Summary**: Combines multiple features to represent interaction effects explicitly - **Subcategory**: Data and Feature Representation ### [Feature Store](https://designpattern.fyi/machine_learning/feature-store/) - **Summary**: Centralized system for computing, versioning, and serving features consistently - **Subcategory**: Data and Feature Representation ### [Hashed Feature](https://designpattern.fyi/machine_learning/hashed-feature/) - **Summary**: Bounded representation for high-cardinality categorical data using hash functions - **Subcategory**: Data and Feature Representation ### [Heuristic Benchmark](https://designpattern.fyi/machine_learning/heuristic-benchmark/) - **Summary**: Compares model performance against simple non-ML baselines for meaningful evaluation - **Subcategory**: Trust, Evaluation and Responsible AI ### [Hyperparameter Tuning](https://designpattern.fyi/machine_learning/hyperparameter-tuning/) - **Summary**: Systematic search for optimal model hyperparameters - **Subcategory**: Training Process and Optimization ### [Keyed Predictions](https://designpattern.fyi/machine_learning/keyed-predictions/) - **Summary**: Passes identifying keys through serving pipeline to match predictions with inputs - **Subcategory**: Serving and Operational Resilience ### [Microservices ML Platform](https://designpattern.fyi/machine_learning/microservices_ml_platform/) - **Summary**: Decomposes the ML platform into independent services: feature store, experiment tracking, training, model registry, serving gateway, monitoring. - **Subcategory**: MLOps ### [ML Pipeline](https://designpattern.fyi/machine_learning/ml-pipeline/) - **Summary**: Automates the workflow of data ingestion, preprocessing, model training, evaluation, and deployment. - **Subcategory**: MLOps ### [Monolithic MLOps Pipeline](https://designpattern.fyi/machine_learning/monolithic_mlops_pipeline/) - **Summary**: All stages of the ML lifecycle run in a single orchestrated pipeline managed by tools like Airflow or Prefect. - **Subcategory**: MLOps ### [Multilabel](https://designpattern.fyi/machine_learning/multilabel/) - **Summary**: Classification where examples can belong to multiple classes simultaneously - **Subcategory**: Problem Framing and Model Structure ### [Multimodal Input](https://designpattern.fyi/machine_learning/multimodal-input/) - **Summary**: Combines different data types (images, text, tabular) into a unified model - **Subcategory**: Data and Feature Representation ### [Neutral Class](https://designpattern.fyi/machine_learning/neutral-class/) - **Summary**: Adds an explicit uncertain class for genuinely ambiguous cases - **Subcategory**: Problem Framing and Model Structure ### [Rebalancing](https://designpattern.fyi/machine_learning/rebalancing/) - **Summary**: Techniques for ensuring models pay adequate attention to rare classes during training - **Subcategory**: Problem Framing and Model Structure ### [Reframing](https://designpattern.fyi/machine_learning/reframing/) - **Summary**: Changes how the ML problem is expressed, often switching between regression and classification - **Subcategory**: Problem Framing and Model Structure ### [Repeatable Sampling](https://designpattern.fyi/machine_learning/repeatable-sampling/) - **Summary**: Deterministic, reproducible train/validation/test splits that prevent data leakage - **Subcategory**: Training Process and Optimization ### [Stateless Serving Function](https://designpattern.fyi/machine_learning/stateless-serving-function/) - **Summary**: Exports models as pure stateless functions for production serving - **Subcategory**: Serving and Operational Resilience ### [Transfer Learning](https://designpattern.fyi/machine_learning/transfer-learning/) - **Summary**: Reuses pretrained model representations for new tasks with less data - **Subcategory**: Training Process and Optimization ### [Transform](https://designpattern.fyi/machine_learning/transform/) - **Summary**: Ensures identical feature transformation logic at training and serving time - **Subcategory**: Data and Feature Representation ### [Two-Phase Predictions](https://designpattern.fyi/machine_learning/two-phase-predictions/) - **Summary**: Splits inference into fast local model and heavier cloud model for efficiency - **Subcategory**: Serving and Operational Resilience ### [Useful Overfitting](https://designpattern.fyi/machine_learning/useful-overfitting/) - **Summary**: Deliberate overfitting when the goal is to approximate a known, deterministic function - **Subcategory**: Training Process and Optimization ### [Windowed Inference](https://designpattern.fyi/machine_learning/windowed-inference/) - **Summary**: Externalizes time-dependent feature computation into stream processing - **Subcategory**: Serving and Operational Resilience ### [Workflow Pipeline](https://designpattern.fyi/machine_learning/workflow-pipeline/) - **Summary**: Orchestrates ML process as discrete, executable components with clear dependencies - **Subcategory**: Serving and Operational Resilience --- ## Machine-Readable Resources - **JSON Feed**: https://www.designpattern.fyi/api/patterns.json - **OpenAPI 3.1**: https://www.designpattern.fyi/api/openapi.json - **llms.txt**: https://www.designpattern.fyi/llms.txt