Data and backend

Python

Where the model, data, and document libraries live.

made by
Python Software Foundation

Magic Ship is one shop in Vancouver, BC, working remotely with clients worldwide. We are not a partner, reseller, or certified vendor of Python Software Foundation - we just build with this.

What Python is

Python is a general-purpose language and the default one for machine learning work. Provider SDKs, evaluation tooling, document parsers, and scientific libraries land here first, and type hints plus Pydantic give enough structure for a production service.

How we use it

Ingestion pipelines, evaluation harnesses, and model-facing services are Python here, usually behind FastAPI with a lockfile pinning the environment. Type hints are not optional in code we hand over, and a type checker runs in CI alongside the tests. Anything with a browser-facing surface is TypeScript instead, and the two sides talk over a client generated from the OpenAPI schema.

Where it is the wrong choice

Two languages in one repo is a real cost - two toolchains, two CI paths, two sets of dependencies to patch - so we do not add Python to a TypeScript codebase unless a library only exists here. Packaging and native dependencies also make Python deploys heavier than the Node equivalent.

Building something on Python?

Send the problem rather than a job spec. You get an answer on scope, on fit, and on whetherPython is even the right call for it.

Start a project

Python and Python Software Foundation are trademarks of their respective owners, used here to say what we work with.