Skip to content
← Back to projects

Autonomous Media System

BlinkyOwl

Production-grade system that generates, publishes, and self-improves educational YouTube Shorts for kids. Private repo, but with public proof via channel and observability dashboard.

2026 PythonRemotionYouTube APIObservabilitySelf-Learning
4 safety gatesDaily autonomous runLive dashboard

The problem

Educational video creation is repetitive, expensive, and hard to sustain consistently. A single short normally means ideation, scripting, illustration, voice, edit, upload, and analytics review. Most creators burn out before consistency ever compounds.

What I built

BlinkyOwl is an autonomous pipeline that handles the entire loop: topic selection, narration, visual planning, audio, rendering, upload, analytics sync, and self-healing adjustments for the next run. It is not just a content generator; it is a monitored production system with gates.

Why it matters

This project is one of the clearest examples of systems thinking in the account. It combines:

  • multi-step orchestration
  • strict kids-content safety gates
  • observability and post-run telemetry
  • feedback loops that improve future runs without letting performance override safety

Public proof despite private code

The repo is private, but the work is still legible because the product output is public:

  • live YouTube channel
  • live observability dashboard
  • concrete operational architecture and quality controls

More Projects

See what else I've built

Hiring tools, prompt governance, automation systems, and open-source utilities.