Boris Brodski · Django · since 2011
Python & Django
Python since 2011 for everything from Django sites to native extensions, machine-learning pipelines and Blender tooling — with the operations side (containers, backups, hosting) done, not delegated.
Django in production. Built, hosted, backed up and maintained — Django and PostgreSQL in Podman.
Python below the surface. Native extensions and a Python ↔ C++ bridge across interpreter versions and 32/64-bit builds.
Python for data and models. TensorFlow, NumPy, Numba, OpenCV — pipelines that train and serve real models.
Tooling and automation. Build and configuration tools, Blender Python API, scripts that run for years.
Web applications
Django came after years of Rails and Java EE, so the architecture habits were already there: clear models, thin views, migrations under control, tests. The current Django/PostgreSQL site is one I operate as well as develop — containers, backups, upgrades.
Native Python
For a CAD manufacturer I built the bridge that lets Python load and process data through a C++ library, across several Python versions and both 32- and 64-bit architectures, with an automated test matrix. The same C++ and build-system background covers packaging, wheels and the parts of Python that are not Python.
Data and models
Python is also the language of my machine-learning work — see AI & machine learning: TensorFlow pipelines, OpenCV inference services, a Gtk desktop application, PostgreSQL feature stores for trading models.
What I do for clients
- Django applications: design, build, deploy, operate.
- Python integration work: C/C++ extensions, bridges, performance (Numba, CUDA).
- Data pipelines and model serving in Python.
Relevant work
-
Web presence for a music quartet
Django/PostgreSQL site built and operated end to end — development, hosting, backups, upgrades.
PythonDjangoPostgreSQL
-
COVID mask detector — vision model from data collection to customer app
Python end to end — training pipeline, OpenCV inference service and the Gtk desktop application.
- Own dataset collected and labelled; model designed and trained from scratch
- Real-time inference on commodity hardware with OpenCV
- Customer-facing Gtk configuration app and web monitoring
PythonTensorFlowOpenCVGtkReactNode.js
-
Python ↔ C++ bridge for a CAD manufacturer
Native extension work — exposing a C++ library to Python across versions and architectures.
PythonC++CPython API
-
AI-assisted cryptocurrency trading
Python data pipeline and PostgreSQL feature store for the trading models.
PythonTensorFlowCUDAGoogle ColabPostgreSQL
Skills in this direction
AI & machine learning
- ★★★★★ Python 2011–
Java
- ★★★★★ Test-driven development 2011–
Web
- ★★★★☆ Django 2022–
- ★★★★☆ PostgreSQL 2006–
- ★★★★☆ MySQL 2002–
- ★★★★☆ HTML / CSS / JavaScript 1998–
- ★★★★☆ React / React Native 2020–
- ★★★☆☆ Node.js 2021–
- ★★★☆☆ OpenAPI / Swagger 2021–
Infrastructure
- ★★★★★ Linux (Debian/Ubuntu) 2004–
- ★★★★☆ Docker 2018–
- ★★★★☆ Podman 2022–
- ★★★★☆ Kubernetes 2024–
Systems & embedded
- ★★★★☆ Blender & Blender Python API 2012–