Shota P.

I build AI video pipelines that run in production: script, voice, avatar, captions, FFmpeg assembly and auto-publishing as one automated system. My own platform, richshorts.lol, is live. Sign up and generate a video to see the pipeline yourself. On Upwork I have delivered three video pipeline contracts: a script-to-finished-video pipeline, a distribution system with FFmpeg, BullMQ, social publishing and RTMP simulcast, and a takeover of an existing Next.js, Postgres and FFmpeg production pipeline. I also built an AI clip generation service on Railway and Supabase with Whisper transcription. What I build and fix: - Topic or script to finished video, end to end - Avatar consistency and lip sync (HeyGen, InfiniteTalk on RunPod, ElevenLabs and MiniMax voices) - Render speed, job queues, retries and per-video cost control - Remotion and FFmpeg assembly with word-level captions Background: 7+ years of enterprise development. I am the lead GIS developer on Georgia's national geospatial platform (nsdi.gov.ge, maps.gov.ge) under a World Bank contract, and I built napr.gov.ge, the public registry portal behind 300+ government services. The systems I ship are built to be maintained, not just demoed. Stack: Node.js, NestJS, TypeScript, Python, PostgreSQL, Supabase, Redis/BullMQ, Docker, AWS Lambda, Angular, Vue, React.