Multi-camera synchronised video publishing platform

Problem

The client had only a concept and sketches for multi-angle video publishing, no brand or design, and needed a working MVP on infrastructure it owned to show investors.

Solution

The team wrote a stack evaluation reviewed by the client's CTO (React on the web, Django and PostgreSQL behind it), agreed a 22-day estimate with a 10-day buffer, and designed a Figma system with a placeholder brand a week ahead of development. The build covered token-authenticated accounts, a home feed with orientation-aware carousels, drag-and-drop multi-upload with reordering and thumbnails, and a player that runs up to four angles under one control, with server-side audio extraction and audio-video merging plus preloading for buffer-free seeking. It was Dockerized and deployed on EC2 behind NGINX with DNS on the client's domain, and the instance was right-sized when merging exhausted CPU credits.

Key features

• Upload of up to 20 synchronised camera angles per take
• Viewer switching between angles mid-playback with audio kept in sync
• Parallel playback of up to four angles under one control
• Drag-and-drop multi-upload with reordering, thumbnails and primary-audio selection
• Server-side audio extraction and audio-video merging
• Buffer-free seeking through preloading
• Free and paid upload limits and video search
• Figma design system with a placeholder brand identity

Tech stack

• AWS EC2
• AWS ECR
• AWS Route53
• Django
• Docker
• Figma
• Formik
• GitHub
• HTML5 video
• Material UI
• NGINX
• Namecheap
• PostgreSQL
• Python
• React
• Redux
• Stripe
• Token authentication
• audio-extract