PROJECT / 004
Video ingestion platform
A cloud-based platform for ingesting and monitoring live camera streams, supporting fall and neglect detection through automated event processing.
- Role
- Software engineer
- System
- Cloud video ingestion and monitoring platform
- Period
- Mar - Apr 2026
Challenge
Keep live video moving reliably.
Build dependable cloud infrastructure and an ingestion path capable of receiving, processing, and monitoring live camera streams while connecting cleanly with the wider backend.
Built
From camera stream to backend event.
Infrastructure and ingestion services for provisioning and deploying DigitalOcean servers, containerizing the platform, processing RTMP streams, and integrating stream workflows with Supabase-backed services.
Technical decisions
Repeatable infrastructure for continuous streams.
- Automated server provisioning with cloud-init and Bash so new Ubuntu hosts could be prepared consistently.
- Containerized the ingestion stack with Docker and Docker Compose to keep deployment and service boundaries repeatable.
- Combined SRS, FFmpeg, and Nginx to receive, process, and route RTMP streams through the platform.
- Connected the ingestion pipeline with Supabase and PostgreSQL services so stream state and downstream event workflows remained coordinated.
Disclosure
The product name, client details, and confidential detection logic are intentionally omitted. This record covers only the infrastructure and ingestion responsibilities that can be discussed publicly.