Skip to content
Book a free 30-minute call with an engineer

Python development

Web backends, data pipelines and AI features

DjangoFastAPIPydanticSQLAlchemyCeleryPandasAirflowpytestuvRuff

Backend

Building with Python

We use Python when a product depends on data, automation or machine learning. That covers web backends, APIs, nightly import jobs and the LLM features clients keep asking for. Almost every useful ML library ships for Python first, so it's the obvious home for that work.

For full platforms with an admin panel we use Django. For lean async APIs and model serving we use FastAPI. Either way the code gets type hints, Ruff, pytest and a Docker image from day one. Python is slower than Go for raw CPU work, and we'll tell you if that matters for your case.

reports.pyPython
1@dataclass(frozen=True)​
2class DailyTotal:​
3 day: date​
4 orders: int​
5 revenue_cents: int​
6​
7​
8def summarise(rows: list[dict]) -> list[DailyTotal]:​
9 by_day: dict[date, list[dict]] = {}​
10 for row in rows:​
11 by_day.setdefault(row["day"], []).append(row)​
12 return [​
13 DailyTotal(d, len(r), sum(x["total"] for x in r))​
14 for d, r in sorted(by_day.items())​
15 ]​

The kind of Python code our team writes every week.

What we build with Python

  1. 01

    Web application backends

    Django or FastAPI backends with login, permissions and an admin your ops team can use.

  2. 02

    APIs & microservices

    Async REST and GraphQL APIs for web apps, mobile apps and partner integrations.

  3. 03

    Data engineering

    Import jobs, scheduled reports and clean-up pipelines with Pandas, Polars, Airflow and plain SQL.

  4. 04

    AI & ML integration

    Model serving, embeddings, retrieval pipelines and LLM features running in production, with cost tracking.

  5. 05

    Automation & scripting

    Back-office scripts, document processing and glue code between systems that currently rely on copy and paste.

  6. 06

    Python 2 & legacy upgrades

    Moving old Python code to a current version, adding tests and type hints along the way.

Why teams pick Python

01

Home of data and AI tools

NumPy, Pandas, PyTorch and the main LLM SDKs all release for Python first.

02

Readable

New developers pick up a well-kept Python codebase quickly, which helps when your team takes it over.

03

Two good web frameworks

Django covers admin-heavy platforms. FastAPI covers small fast APIs. We use both.

04

Types when you want them

Type hints checked by mypy, plus Pydantic models, catch a lot of bugs in large codebases.

Good fits

  • AI features inside SaaS products
  • Analytics and reporting platforms
  • Data pipelines
  • Admin-heavy business systems
  • Automation tools

Where we use it

Python questions

Is Python fast enough for a production web app?

For most web apps, yes. Response time is mostly spent waiting on the database and the network, not on Python itself. FastAPI handles lots of concurrent requests well, and caching and decent queries matter more than the language. If one part really is CPU-bound, we use a compiled library or move that piece into its own service.

Should we use Django or FastAPI?

Django if you want an ORM, migrations, an admin panel and auth out of the box. FastAPI if you need a lean async API, a microservice or model serving. Some of our projects use both: Django for the main product and a FastAPI service for the AI or high-traffic parts.

Can you add AI features to our existing product with Python?

Yes. Typical requests are semantic search, document summaries, a support assistant or recommendations. We handle the data prep, retrieval, prompts, evaluation and cost limits, then expose it to your app through a normal API so the rest of your stack doesn't change.

How do you deploy Python applications?

We build Docker images and run them on AWS ECS, Google Cloud Run, Azure App Service or Kubernetes for bigger systems. The CI pipeline runs tests, linting and a security scan before every release, and monitoring and alerts are set up before launch.

Pairs well with