ProfessionalBrilworks SoftwareFull-stack developer2024 — 2026
Video AI Learning System
HLS learning platform with a serverless video pipeline.
A learning platform with a custom HLS player that resumes where students left off, and an automated AWS Lambda + FFmpeg pipeline that packages every upload for streaming.
0 manual
steps from upload to stream
01
Why I built it
Course videos were processed by hand before students could watch them.
02
The problem
- Manual transcoding.
- Students losing their place across sessions.
03
System design
- 1
Upload to S3
Triggers the processing pipeline.
- 2
AWS Lambda + FFmpeg
H.264 remux and HLS packaging.
- 3
HLS player
Resume playback from the last watched position.
04
Decisions & trade-offs
Serverless video processing
Why: Uploads are processed automatically and the cost follows usage.
Trade-off: Lambda time and memory limits shape how large a video can be.
05
Backend & frontend
Backend
- Lambda + FFmpeg ingestion, S3 storage and playback URLs.
Frontend
- Custom HLS player with resume playback.
06
Results
- Videos go from upload to streaming without manual steps.