Techniques for the automated customization and deployment of a machine learning application

1. A method for automatically creating a machine-learning application for use in a production environment, the method comprising:

receiving an input, wherein the input identifies a problem to be solved using the machine-learning application;

selecting a machine-learning model template from a plurality of templates based at least in part on the input, wherein the machine-learning model template includes metadata, the metadata specifies data expectations and available data formats;

analyzing one or more formats of customer data to generate a customer data schema based at least in part on a data ontology that applies to the identified problem;

determining whether the customer data schema is misaligned with one or more key features of the selected machine-learning model template;

based at least in part on the determination that the customer data schema is misaligned with the one or more key features of the selected model, analyzing the metadata for the selected machine-learning model template to determine what additional information is required to re-align the customer data with the data expectations of the selected machine-learning model template as indicated in the metadata;

gathering the additional information required to re-align the customer data with the data expectations of the selected machine-learning model template;

based at least in part on the additional information, performing one or more remedial actions on the customer data to re-align with the data expectations of the selected machine-learning model template to result in a transformed dataset, wherein a remedial action of the one or more remedial actions comprises rescaling the customer data to fit the data expectations of the selected machine-learning model template, the rescaling of the customer data includes an infrastructure deployment corresponding to one or more micro services architecture in the transformed dataset; and

training the machine-learning application using the selected machine-learning model template and the transformed dataset.

2. The method of claim 1 , wherein the metadata for the selected machine-learning model template is derived based at least in part on historical model inputs.

3. The method of claim 1 , wherein the determining whether the customer data schema is misaligned comprises determining that the customer data is entirely missing one or more features that were expected.

4. The method of claim 1 , wherein the determining whether the customer data schema is misaligned comprises determining that the customer data contains values for one or more features, and determining that the customer data does not contain enough values for the one or more features to satisfy the data expectations of the selected machine-learning model template; and

further comprising adding, to the customer data, duplicates of one or more of the values for the one or more features.

5. The method of claim 1 , wherein the determining whether the customer data schema is misaligned comprises determining that one or more units of the customer data is different from the one or more key features of the selected machine-learning model template.

6. The method of claim 1 , further comprising:

determining that the customer data is out of scale with the data expectations of the selected machine-learning model template; and

based at least in part on the determining, executing the remedial action comprising the rescaling of the customer data to fit the data expectations of the selected machine-learning model template.

7. The method of claim 1 , wherein the determining whether the customer data schema is misaligned comprises determining that a data distribution of the customer data is different from the data expectations of the selected machine-learning model template based at least in part on historical data distributions for different applications of the machine-learning model template.

8. The method of claim 1 , wherein the determining whether the customer data schema is misaligned comprises determining that a data distribution of the customer data is different from the data expectations of the selected machine-learning model template based at least in part on detecting a lack of data continuity with respect to at least one metric.

9. The method of claim 1 , wherein the determining whether the customer data schema is misaligned comprises determining that a data distribution of the customer data is different from the data expectations of the selected machine-learning model template based at least in part on detecting a lack of data coverage for one metric based at least in part on a particular value of at least one other metric.

10. A server system, comprising:

one or more memories storing instructions; and

one or more processors communicatively coupled to the one or more memories to execute the instructions to perform operations to:

receive an input, wherein the input identifies a problem to be solved using a machine-learning application;

select a machine-learning model template from a plurality of templates based at least in part on the input, wherein the machine-learning model template includes metadata, the metadata specifies data expectations and available data formats;

analyze one or more formats of customer data to generate a customer data schema based at least in part on a data ontology that applies to the identified problem;

determine whether the customer data schema is misaligned with one or more key features of the selected machine-learning model template;

based at least in part on the determination that the customer data schema is misaligned with the one or more key features of the selected model, analyze the metadata for the selected machine-learning model template to determine what additional information is required to re-align the customer data with the data expectations of the selected machine-learning model template as indicated in the metadata;

gather the additional information required to re-align the customer data with the data expectations of the selected machine-learning model template;

based at least in part on the additional information, performing one or more remedial actions on the customer data to re-align with the data expectations of the selected machine-learning model template to result in a transformed dataset, wherein a remedial action of the one or more remedial actions comprises rescaling the customer data to fit the data expectations of the selected machine-learning model template, the rescaling of the customer data includes an infrastructure deployment corresponding to one or more micro services architecture in the transformed dataset; and

train the machine-learning application using the selected machine-learning model template and the transformed dataset.

11. The server system of claim 10 , wherein the metadata for the selected machine-learning model template is derived based at least in part on historical model inputs.

12. The server system of claim 10 , wherein the determination whether the customer data schema is misaligned comprises determining that the customer data is entirely missing one or more features that were expected.

13. The server system of claim 10 , wherein the determination whether the customer data schema is misaligned comprises determining that the customer data contains values for one or more features, and determining that the customer data does not contain enough values for the one or more features to satisfy the data expectations of the selected machine-learning model template; and

further comprising adding, to the customer data, duplicates of one or more of the values for the one or more features.

14. The server system of claim 10 , wherein the determination whether the customer data schema is misaligned comprises determining that one or more units of the customer data is different from the one or more key features of the selected machine-learning model template.

15. The server system of claim 10 , wherein the one or more processors is configured to:

determine that the customer data is out of scale with the data expectations of the selected machine-learning model template; and

based at least in part on the determination, execute the remedial action comprising the rescaling of the customer data to fit the data expectations of the selected machine-learning model template.

16. The server system of claim 10 , wherein the determination whether the customer data schema is misaligned comprises determining that a data distribution of the customer data is different from the data expectations of the selected machine-learning model template based at least in part on historical data distributions for different applications of the machine-learning model template.

17. The server system of claim 10 , wherein the determination whether the customer data schema is misaligned comprises determining that a data distribution of the customer data is different from the data expectations of the selected machine-learning model template based at least in part on detecting a lack of data continuity with respect to at least one metric.

18. The server system of claim 10 , wherein the determination whether the customer data schema is misaligned comprises determining that a data distribution of the customer data is different from the data expectations of the selected machine-learning model template based at least in part on detecting a lack of data coverage for one metric based at least in part on a particular value of at least one other metric.

19. A non-transitory computer-readable medium storing instructions, the instructions comprising:

one or more instructions that, when executed by one or more processors, cause the one or more processors to:

receive an input, wherein the input identifies a problem to be solved using a machine-learning application;

select a machine-learning model template from a plurality of templates based at least in part on the input, wherein the machine-learning model template includes metadata, the metadata specifies data expectations and available data formats;

analyze one or more formats of customer data to generate a customer data schema based at least in part on a data ontology that applies to the identified problem;

determine whether the customer data schema is misaligned with one or more key features of the selected machine-learning model template;

based at least in part that on the determination that the customer data schema is misaligned with the one or more key features of the selected model, analyze the metadata for the selected machine-learning model template to determine what additional information is required to re-align the customer data with the data expectations of the selected machine-learning model template as indicated in the metadata;

gather the additional information required to re-align the customer data with the data expectations of the selected machine-learning model template;

based at least in part on the additional information, perform one or more remedial actions on the customer data to re-align with the data expectations of the selected machine-learning model template to result in a transformed dataset, wherein a remedial action comprises rescaling the customer data to fit the data expectations of the selected machine-learning model template, the rescaling of the customer data includes an infrastructure deployment corresponding to one or more micro services architecture in the transformed dataset; and

train the machine-learning application using the selected machine-learning model template and the transformed dataset.

20. The non-transitory computer-readable medium of claim 19 , wherein the metadata for the selected machine-learning model template is derived based at least in part on historical model inputs.