polyform 0.2.9

Serverless DataOps, for AI/ML purposes (Prototype)

Polyform

Proof of Concept

Polyforms are functions-as-a-service which include DataOps, specifically oriented for running AI systems:

• are designed for native AI functionality
• have an explicit expectation of the shape of data sent to them
• support different sub-forms for variant purposes (such as training, vs intersect)
• support transformation of data in flight, as well as storing back to the data universe
• come with the implicit security of the Data Multiverse

Data Expectations describe the shape of the data as needed, and draw it in on demand.

This code is a combination of a very rough implementation (prototype) to prove the concept, combined with CI/CD automation for ease of use.

It is targeted towards the data science industry, enabling deep learning to be more easily managed as production ready services.

pip3 install 'git+ssh://[e-mail]/srevenant/polyform.git@master#egg=polyform'

Run poly with no args for a breakdown of options

Tutorial coming soon