How we use Python
Releasing apps in several languages involves many small files: store descriptions, screenshots at different sizes, translated strings. We automate that pipeline in Python, so assets are generated and checked the same way for every release rather than assembled by hand.
For backends, FastAPI gives typed request validation and generated API documentation with little code. Python is also where we analyse aggregated, anonymous product metrics and prototype an idea before deciding which language the production version should use.
What we build with it
- REST APIs with FastAPI
- Store asset and localisation pipelines
- Data processing and reporting scripts
- Prototypes for new features and algorithms
Toolkit
- FastAPI
- pandas
- Pillow
- pytest
- Poetry
- Jupyter
How we keep Python code healthy
- Type hints checked with mypy
- Formatting and linting with Ruff
- Tests with pytest in CI
- Dependencies pinned per project
Questions about Python
What do you automate with Python?
Mostly release chores: screenshots, store texts, string files and checks that nothing is missing before upload.
Is Python fast enough for a backend?
For most app APIs, yes. Where throughput matters most, we write that service in Go.