IJ Connect Improving a platform people use every day.
Production Django work at ITSS Paris: faster reporting, private messaging, and full-stack delivery for approximately 5,000 professionals across France.
Software Engineer · ITSS Paris · August 2025–present
My work on IJ Connect combines feature delivery with production support: understanding a requirement, implementing it across the stack, validating the release, and supporting the people who use it.
The production context
IJ Connect is a Django platform used daily by approximately 5,000 professionals across France. As a Software Engineer at ITSS Paris, I contribute features, improvements, and support to this existing production system.
From minutes to about 30 seconds
Statistics pages took several minutes to load. I traced PostgreSQL bottlenecks back to the Django ORM workflows and designed date-based indexes and range partitioning. The reported page-load time fell to approximately 30 seconds.
The work connected application-level queries with how the database stores and retrieves reporting data. The before-and-after timing applies to the statistics pages involved in this optimization.
Private messaging, beyond the chat box
I integrated Matrix with Django Channels and WebSockets for private messaging and live notifications. The work included controlled rooms, invitations, and permissions, alongside the user-facing interactions.
The engineering scope extended beyond displaying messages: access rules and room participation needed to be handled alongside real-time updates. This public account describes the integration scope without exposing client implementation details.
Delivering across the stack
- Model the feature
- Data models, migrations, business rules, and Django administration tools.
- Connect the interface
- Responsive forms and AJAX endpoints using JavaScript, Django templates, Bootstrap, and LESS.
- Support the release
- Requirements analysis, estimation, deployment support, release validation, and production incident handling, with direct communication about constraints and timelines.
Outcome and boundaries
The work demonstrates how I contribute to a live platform: improve an expensive reporting workflow, deliver real-time functionality with access controls, and support features beyond their first release.
The source code is proprietary. The public launch announcement provides product context; the implementation scope and approximate metrics here come from my professional experience. No client source code, internal dashboards, or private user data are published.