Streamlining AI development and Deployment with KitOps
summary
Claims that 80% of ML models never reach production and shows a concrete path to fix that using ModelKits and KitOps, with live CLI demos unpacking a YOLOv10 kit and repacking and publishing it to an OCI-compliant registry. Commands shown include kit unpack, kit pack and kit push, creating layered artifacts that behave like Docker images. Integration with CI/CD is demonstrated using Dagger and GitHub Actions so model kit versioning, selective unpacking and Kubernetes init containers all work end to end.
AI-written summary · about
chapters (8)
- 0:00Welcome to KitOps
- 1:07AI deployment challenges
- 8:56Introducing KitOps
- 10:46How model kits work
- 14:40Live demo of KitOps
- 22:31Automating ML pipelines
- 27:04Resources and contributions
- 27:48Q&A on Docker and datasets
related talks
23mBringing Container-Native Simplicity to AI/ML
32mBridging DevOps and MLOps: Unifying Pipelines with KitOps and GitOps
24mFrom Model to Production: Automating the ML Lifecycle with CI/CD
21mMastering the DevOps Machine Learning Pipeline
57mDevOps for Machine Learning
23mAccelerating Model Deployment with MLOps
44mDevOps for Machine Learning - why is it different?
26mApplying DevOps practices in Data and ML Engineering