AWS-ML: How to deploy/setup my own ML algorithms on AWS platform as pay-to-use API?

Question

The title sums it up. Essentially, I'd like to offer my own closed-source proprietary ML algorithms to Amazon AWS customers on a pay-to-use basis API - e.g., sales volumes prediction algorithm service licensed monthly or annually or per call. Most information found talks about how to build and give it away, or use it internally within one's company, but I'm looking to offer it to the public as a commercial offering on AWS.

Thank you in advance for your help - links to articles, help pages, or direct steps on how to do this.

Answer 1

Allow me please to answer my own question. Although not a 100% what I was hoping for, there's certainly support for this in the platform which is great to see: Software-as-a-Service-Based Products

Answer 2

This is actually very easy to do with AWS.

Create an AWS Marketplace account.

Create an AMI bundled with your software with per hour pricing.

This link will get you started:

Sell on AWS Marketplace

Answer 3

It looks like you need to set up your own EC2 server and then create an API gateway call to invoke your service, and you can create API keys to control access/pricing for your end users.

API Key usage: https://docs.aws.amazon.com/apigateway/latest/developerguide/api-gateway-setup-api-key-with-restapi.html

You can also look into AWS Sagemaker to set up your ML pipeline and provide a managed inference endpoint if you don't want to host your own EC2 server, from there can leverage API gateway the same way.