Asset Graph

Asset Graph is a powerful system designed to compose a searchable graph of files based on a set of contextual associations. The system is architected as a modular framework with extension points for expanding supported file types and relational mapping techniques, making it highly flexible and customizable to the dataset and application.

The system's capabilities were successfully demonstrated against a dataset from a notable animated YouTube series. Our client successfully indexed 400k+ production assets. The generated graph was then used to aggregate relevant training data for custom generative AI models.