Multilingual captions for AI video calls
A full-stack feature (Django REST backend, React frontend) shipped in hours by pairing with Claude Code on a 500+ file codebase.
A production feature for an AI video and audio calling product: live captions in multiple languages, managed by admins.
What shipped
- Backend: Django REST Framework APIs with full CRUD, Redis caching (30-minute strategy) and PostgreSQL
- Database: migrations that auto-populate sensible defaults
- Frontend: React with Ant Design: language management, drag-and-drop reordering and real-time UI updates
- Admin: multiple languages managed and ordered from the admin panel
How it was built
I worked with Claude Code in the terminal across a codebase of 500+ files. I set the architecture and data flow, described each piece in plain English, and reviewed every change. When a git merge overwrote UI work, the AI analysed what was lost, re-implemented it, and fixed a duplicate-key bug along the way.
Work that would usually take weeks took hours. The architecture and the review are still mine; the AI removes the slow parts.