Automated video pipeline & technical due diligence

The brief

Two engagements in the same domain: developing an automated, AI-driven video-creation pipeline, and giving a Swiss investor an honest technical read on a short-form video product they were backing.

The AI video-creation pipeline

We developed and hardened an automated pipeline that turns a brief into a finished video — built in Python with FastAPI and a Celery task queue, behind a React / tRPC frontend. The work spanned the full chain: a scriptwriting agent, automated stock-footage fetching, brand-compliance checks, video rendering, and a character-lock system that keeps on-screen characters consistent across scenes. It was engineered for production — config-driven, queue-based, and built to scale render workloads rather than fall over on the second job.

Investor technical due diligence

Separately, a Swiss investor needed to know whether a short-form, TikTok-style video platform they were backing was as far along as claimed. We audited the full codebase — a Node.js / Express backend, a React admin panel, and a Flutter mobile app — assessed real MVP completeness at roughly 45–55%, and delivered an 11-page professional status report covering architecture, gaps, risks, and the work remaining.

The outcome

On the build side: a working, scalable AI video pipeline with its hardest pieces rendering and character consistency — solved.