DatologyAI

DatologyAI, founded in 2023, develops automated tools for curating AI training data. Its platform processes data at petabyte scale, employing model-based filtering, embedding-based filtering, and synthetic data integration to select and optimise datasets for deep learning models. The company reports that this approach yields training speedups of between seven and 40 times.

The platform addresses a core bottleneck in machine learning: the preparation and quality of training data. By automating curation, DatologyAI aims to reduce the manual effort and compute required to train deep learning models. Its technical domains span AI data curation, automated data selection, and synthetic data integration.

DatologyAI operates in the AI development and deep learning sectors. Details on team size, funding, and geographic presence were not disclosed.