A collection of custom Django admin interfaces built for a healthcare management platform. These go well beyond the default admin and into territory where most teams would reach for a separate frontend.

The most complex piece is a questions & answers tree editor for configuring an AI assistant’s decision logic. Each tree is a self-referencing hierarchy: a root question branches into answers, which lead to either result nodes or further questions, nested arbitrarily deep. Built with django-nested-inline, but heavily customized. Out of the box, django-nested-inline only supports drag & drop within the same nesting level, and doesn’t support self-referencing foreign keys at all. I worked around both limitations: you can drag a subtree from any depth and drop it onto any “Add another” link at any other level, with the dragged content collapsing cleanly during the move.

Another piece is a license matrix widget: a single admin page for managing many-to-many relationships across entire model. Each relationship is rendered as a checkbox in a grid, with three tools that work on both rows and columns depending on cursor position: check all, uncheck all, and copy (which applies an entire row or column to another). This replaced the need to manage hundreds of individual M2M relationships across separate admin pages, turning it into a single visual interface editable by non-technical staff.

Also built the admin section for a versioned configuration system for an AI Assistant. The configuration supports exports and imports with validation. The section also includes the ability to configure and run regression test suites against the AI Assistant. Test suites run 200+ cases asyncronously, using several parallel runners. The LLM responses are compared against pre-validated expected outputs, with pass/fail tracking, execution time per case, and an overall statistical score is calculated.

 

Technologies: Django, Celery, AWS (serverless), Vanilla.JS

Client name and identifying details withheld due to confidentiality. Screens and data are recreated/illustrative; no proprietary assets included.