In plain English
An AI project is a loop: define the problem, check the data, prototype, evaluate, deploy, monitor and improve. Write the goal as a number. Models drift as the world changes, so monitoring never stops. MLOps and LLMOps are the habits that keep AI working.
- Drift
- A slow change in data or right answers that makes a model worse.
- MLOps
- Habits and tools for running machine learning in real life.
- Pilot
- A small first release to a few people.
Picture it
The loop that never ends
1
Define and prepare
Goal as a number, good data
2
Build and evaluate