In AI/ML projects, various components are usually stored in separate locations:
- Code resides in Git repositories
- Datasets and models are stored in DVC or storage services like S3
- Parameters are managed using experiment management tools
As components are stored separately, the process of deploying and integrating them can become more complicated.
KitOps’s ModelKits offers a unified solution by packaging these components into ModelKits. This allows for easy versioning and sharing of components with other team members in just a few commands.