KitOps is a CNCF open standards project, and ModelKits are packaged, versioned, and shipped by teams that need their AI/ML projects to move through the same pipelines as everything else they run. These are the organizations building on it.
Teams who have written up how they put KitOps to work, in their own words.
A topological deep learning research lab that standardized on ModelKits as its packaging and versioning standard — bundling the model card, MLflow experiment pointers, LakeFS dataset references, and configuration into a single OCI artifact it promotes from dev to staging to production.
“It felt like discovering containerization. It feels like Docker for models.”
An IT security provider delivering AI/ML projects for government agencies and security-conscious organizations, using ModelKits to package datasets and model milestones across its development lifecycle and to automate Kubernetes deployments.
“We have MLflow, but executives need a tamper proof and auditable solution. We were considering customizing MLflow to add secure storage but now we don’t have to!”
Organizations and projects using KitOps in their AI/ML workflows.