Interpretable AutoML at scale created by and for data scientists
Train data parallelization and leave subsampling behind.
Complete transparency regarding executed algorithms, resulting models and feature analytics for interpretability.
Self-monitoring audit engine to ensure that every execution is aligned with the behavior expected.
Cloud-based AutoML tool to continuously train predictive models by exploiting the freshest data available, reducing algorithmic-expert supervision and deploying to production effectively.
Harnessing extensive open-source ML algorithms, ranging from simpler methods, including logistic regression, to more complex ones, such as XGBoost, CatBoost and LightGBM.
Providing feature-importance rankings, partial dependence plots and human-interpretable explanations obtained from linear surrogate models.
Monitoring tool that controls changes in acceptance rates, score distributions, model decisions and variations in data that define the target population.
