Geo-MLOps Manual
The complete guide to Geo-MLOps — one platform for the whole life of a model, from datasets to field devices.
Geo-MLOps covers the entire life of a machine-learning model on one screen: datasets → training → experiment tracking → model registry → approval → serving and field deployment → monitoring → retraining. This manual is written screenshot-first so that anyone can follow along.
Where to start
| I… | Start here |
|---|---|
| am evaluating the platform | Why MLOps — what changes, in ten minutes |
| install and operate the platform | Installation → Administration |
| manage a tenant's members, tokens and training environment | Administration |
| upload data and train and deploy models | User guide — start with the 30-minute tutorial |
| build training images or connect a DataOps system as a partner | Integration guide |
| connect field (edge) devices | Edge SDK |
| look up APIs, settings or permissions | Reference |
How to read this manual
- The platform UI is in Korean. Button and menu names are shown as English name (한국어 label) so you can find them on screen.
- Red numbers in screenshots (①②③) are explained with the same numbers in the text. Click a screenshot to enlarge it.
- The 🔑 badge at the top of a page is the permission you need for that feature. See Roles and permissions.
- The tenant
DEMO, accounts at@example.comand the hostmlops.example.comare placeholders — substitute your own. - Use 한국어 at the top right for the Korean original, and Ctrl+K to search.