← All work
ArchitectureCloud EngineeringCloudBackend
Serverless YouTube File Sync
Event-driven AWS pipeline that discovers new YouTube content, queues downloads, stores files in S3, and prevents duplicate processing through DynamoDB.
github · source
Architecture/ the problem
Syncing and processing YouTube content on a schedule needs an event-driven architecture that stays cheap at idle, retries safely, and never double-processes the same video.
/ approach
TrustCode designed the serverless AWS pipeline, scheduling, queuing, idempotent state, and cost-aware storage.
/ what we built
- AWS CDK (TypeScript) infrastructure-as-code for the full stack
- EventBridge hourly schedule → search Lambda → SQS → download Lambda
- Custom arm64 Docker runtime with yt-dlp and ffmpeg
- DynamoDB idempotency, S3 storage with lifecycle expiry, CloudWatch metrics, and DLQ retries
/ the outcome
A documented serverless pipeline architecture, scheduled discovery, queued downloads, idempotent state, and cost-aware storage controls.
/ next case study
Multi-Carrier Shipping Integration