Jochen Wersdörfer

Some teams face constraints that make standard web deployment impossible: confidential data that must never leave user machines, air-gapped environments, and workloads that are too expensive to run centrally. This talk presents a production pattern for turning Django into a desktop application by packaging it inside Electron. The case study is Steel-IQ, an open-source steel-industry simulation tool for multi-decade decarbonization scenarios, shipped for local execution with sensitive data. Structure (30 minutes): 1. Why this architecture exists: confidentiality, offline environments, heavy compute. 2. System design: Electron + Django + django-tasks workers + SQLite (including WAL mode). 3. Packaging and distribution: python-build-standalone, uv, installers, and update paths via GitHub Actions. 4. Production failure modes: startup checks, graceful shutdown, orphan process cleanup, migrations, and memory behavior. 5. Demonstration and decision checklist: when this approach is right, and when to choose alternatives. Audience and level: - Intermediate to advanced Django developers. - Useful for teams building internal tools, offline-first products, or applications with strict data constraints. - No prior Electron experience required. Attendee takeaways: 1. A reference architecture for embedding Django inside Electron in production. 2. A concrete packaging strategy for standalone Python across platforms. 3. Process-management patterns for reliability and recovery. 4. Operational lessons for long-running local workloads. 5. A decision framework for when desktop-packaged Django is a fit.