Library · Python / Software Engineering
AutomationML
A library for assembling ML pipelines and searching hyperparameters.
- Python
- scikit-learn
- Optuna
Project and contribution
Developing shared model interfaces, a registry and factory, parameter validation and scikit-learn pipeline assembly. Adding GridSearch, RandomizedSearch and Optuna search wrappers, tests and CI configuration.
Purpose
The package is available on TestPyPI.
AutomationML brings recurring machine-learning steps together: model creation, configuration validation, pipeline assembly and hyperparameter search. It exposes these operations through shared interfaces.
My work
- Develop base interfaces, a model registry and a factory.
- Add lazy dependency loading and parameter validation.
- Combine preprocessing and models into scikit-learn pipelines.
- Implement search wrappers for GridSearch, RandomizedSearch and Optuna.
- Maintain tests and automated-check configuration.
Technical approach
The library provides a common layer over existing ML tools. External libraries supply the models and algorithms; my work covers their interfaces, integration and configuration checks. The project is licensed under Apache 2.0.