Automating ML Pipelines with AutoGluon, Leading Frameworks, and Real-World Integration
Complete code companion with production-ready Jupyter notebooks covering feature engineering, neural architecture search, tabular data, NLP, time series, computer vision, and MLOps deployment.
Production-ready code implementations for every major topic
Master data preprocessing with intelligent profiling, outlier detection, encoding strategies, and multimodal feature fusion using a simulated e-commerce dataset.
Implement NAS from scratch with search space design, evolutionary algorithms, DARTS, performance estimation strategies, and one-shot supernet training.
Complete guide to AutoGluon's TabularPredictor for classification and regression, from basics to advanced customization with SHAP interpretability.
Deep dive into text classification, NER, semantic matching, and multimodal text+tabular processing with AutoGluon's MultiModalPredictor.
Build forecasting systems with AutoGluon including Chronos-Bolt foundation models, probabilistic calibration, and business impact optimization.
Retail product classification, multimodal image+text learning, CNNs vs Vision Transformers, object detection, and production deployment patterns.
Integrate AutoML into production with MLflow experiment tracking, model registry, Kubeflow pipelines, drift detection, and automated retraining.
Build robust data pipelines with validation gates, point-in-time feature engineering, incremental processing, and data contracts.
Production deployment patterns with FastAPI, shadow deployments, Prometheus metrics, Evidently drift detection, and continuous learning.
End-to-end production implementations with real business impact
Production-ready implementation processing 50M daily transactions with 62% reduction in fraud losses and <100ms latency.
Omnichannel forecasting across 450 stores × 50K SKUs with stockout detection, weather effects, and hierarchical predictions.
Fairness-aware AutoML for 30-day readmission risk with HIPAA compliance, multimodal EHR learning, and clinical interpretability.
Set up your environment and start learning in minutes
git clone https://github.com/tomakk/OReillyAutoMLBook.git
conda create -n automl python=3.10 && conda activate automl
pip install autogluon torch tensorflow mlflow optuna
jupyter lab