CASE STUDY 04 · MOBILE / MARKETPLACE
Livelihood Mobile App
A render-to-publish pipeline that turns a script into a posted video with no manual export-and-upload step in between.
ROLE
Backend + automation engineering
STACK
NestJS · BullMQ · AWS
TIMELINE
[x weeks]
TYPE
Client project
01 — PROBLEM
Every short video meant exporting from an editor, renaming files, and posting to each platform by hand — ten to fifteen minutes of repetitive work per clip, for creators publishing daily.
02 — ARCHITECTURE
Script / assets
input
Render queue
BullMQ
Workers
render job
Platform APIs
social posting
Published post
scheduled or live
03 — KEY DECISIONS & TRADE-OFFS
Queue renders instead of rendering in-requestA render can take minutes; BullMQ workers pick up jobs, retry failures and report progress without holding a request open.TRADE-OFF Needs its own worker infrastructure and a job-status UI, where a simple script had neither.
One pipeline, many platform targetsA render job fans out to each platform's own posting API, so one render produces every version a creator needs.TRADE-OFF Each platform's API has its own quirks and rate limits to handle separately.
Schedule posts, don't just publish immediatelyCreators plan posting times in advance; the pipeline queues a post for its scheduled time instead of always publishing the moment a render finishes.TRADE-OFF Needs reliable scheduled-job execution, not just a one-shot queue.
04 — STACK
NestJS
Pipeline orchestration
Queue workers
Render jobs
Social APIs
Multi-platform posting
AWS
Render compute & storage
05 — RESULTS
[metric]
Minutes of manual work saved per clip.
[metric]
Clips published per week, same team size.
[metric]
Posting errors needing a manual retry.
NEXT CASE STUDYBehold Bohol Trip →