Back to Portfolio
MEDIA AUTOMATION

Creator-First Video Editing Automation

Targeted ML on audio and vision cutting 80% of manual editing time for short-form creators

Challenge

  • Creators spent hours manually scanning videos for 'interesting' moments
  • Repetitive cutting, trimming, and aligning to audio peaks
  • Off-the-shelf editors didn't optimise for fast, template-driven social content

Solution

  • Built a video editing automation engine that uses ML peak sound detection to find high-energy, high-engagement segments
  • Applies computer vision to detect scene changes and key visual events
  • Proposes an auto-edited timeline aligned to audio & visual cues
  • Developed a Vue.js frontend with simple upload experience and visual timeline preview
  • Controls for templates and export presets tailored to social platforms
  • Implemented a Python + Flask backend to handle video processing workloads, orchestrating CV/audio ML steps
  • Manage storage, job status, and exports
  • Deployed on DigitalOcean, tuned for cost-efficient scaling for a small startup

Impact

  • Dramatically reduced manual selection and cutting for early users
  • Created a solid proof-of-concept for AI-assisted editing geared to social media
  • Demonstrated ability to go from idea → ML prototype → deployed app

Tech Stack

Frontend

  • Vue.js
  • JavaScript

Backend

  • Python
  • Flask
  • Video/audio ML
  • Computer vision

Infrastructure

  • DigitalOcean
  • Containerised deployment

Ready for AI that drives revenue?

HEXALON