DatologyAI is a 2023-founded company that builds automated tools to select and optimise training data for deep learning models. Its platform processes data at petabyte scale, using techniques such as model-based and embedding-based filtering along with synthetic data integration to identify high-value data points and remove redundant or misleading information. The company's stated goal is to help organisations train models faster, improve performance, and cut computational costs.
The firm's automated data curation platform has demonstrated training speedups of 7-40x, and its approach allows smaller models to outperform larger ones trained on uncurated data. The team draws from experience at leading AI research labs, bringing expertise in AI data curation and deep learning to the problem of data quality at scale.
DatologyAI operates in the AI development vertical. Founded in 2023, the company's funding details and team size have not been publicly disclosed.
