MLOps Pipeline
Data flows into a feature store, a training job writes to a model registry, serving pulls the model, and monitoring closes the retrain loop.
Rendering…
Make it your own.
{
"nodes": [
{ "id":"data", "type":"cylinder", "label":"Data", "x":40, "y":170, "fill":"#fef3c7" },
{ "id":"feat", "type":"vector-db", "label":"Feature store", "x":260, "y":160, "fill":"#a5f3fc" },
{ "id":"train", "type":"gear", "label":"Training", "x":480, "y":160, "fill":"#fde68a" },
{ "id":"reg", "type":"db-cluster", "label":"Registry", "x":700, "y":160, "fill":"#a5f3fc" },
{ "id":"serve", "type":"iso-cube", "label":"Serving", "x":920, "y":160, "fill":"#c7d2fe" },
{ "id":"mon", "type":"donut-3d", "label":"Monitoring", "x":1140, "y":160, "fill":"#dbeafe" }
],
"edges": [
{ "id":"e1","source":"data","target":"feat" },
{ "id":"e2","source":"feat","target":"train" },
{ "id":"e3","source":"train","target":"reg","label":"model" },
{ "id":"e4","source":"reg","target":"serve","label":"deploy" },
{ "id":"e5","source":"serve","target":"mon" },
{ "id":"e6","source":"mon","target":"train","label":"retrain" }
]
}