AI/ML components

Platform documentation defines supported behavior. Implementation detail stays with the owning repositories and upstream projects.

Component	Responsibility	Source baseline
Oqullus	AutoML, Experiments, and Models journeys and state presentation	sparqd/oqullus at `f76e57dbb88d`
BFF API	Authenticated UI facade and MLflow token exchange	sparqd/bff-api at `93ea18e72414`
AutoML backend	Workflow planning, data access, training, status, prediction, and generated artifacts	sparqd/agent-based-ml-platform at `3112ee9f7e72`
Platform stacks	Service identities, deployment, storage, database, capacity, and recovery policy	sparqd/platform-stacks at `5f3f0f0e44fb`
MLflow	Experiment tracking, artifacts, and registered-model metadata	MLflow tracking and Model Registry
AI Gateway	External GenAI provider routing used by AutoML reasoning; not trained-model serving	sparqd/ai-gateway at `94236c69c13d`
Trino and Ranger	Training-data query execution and physical data policy	Deployment-owned runtime services