Creating a Data Machine
Every Data Machine is a workflow of steps executed as a series of operations either based on a task assigned or by validating a condition. Hence, building a Data Machine requires an organization of steps in the order they need to be executed to lead to the final result.
- Assign an AI model to execute a desired task on a user provided input or the result of a previous step.
- Assign an AI model to perform a validation of a result from a previous step or a user provided input
If all the desired Operational and Conditional steps are added, then create on last step called the Final Step that determines how the output of the Data Machine can be utilized.
Every Data Machine requires Testing to be performed by the creator to ensure the right output is being generated. This is a mandatory step for every Data Machine, regardless of which AI models are used.
Data Machines are created in a Draft state initially until explicitly Published. The creator of a Data Machine has the option to test the Data Machine as many number of times as needed before Publishing the Data Machine. Only Published Data Machines are executable using Rest API or any of the other execution options.
