n8n Automation Engineer — AI-Powered Video Repurposing System (Whisper + GPT-5)

I create a high volume of long-form video content — podcasts, interviews, keynotes, and talks — and I need a smart, repeatable system that automatically converts each video into polished, platform-ready short clips without any manual work on my end. I'm looking for an experienced automation developer to build a clean, well-documented n8n workflow I can trigger again and again on new videos. Here is what the workflow needs to do: Accept a long-form video through either a direct file upload or a YouTube/Vimeo URL, feeding it into the pipeline automatically Transcribe the full audio using Whisper with word-level timestamps, so every word has an exact time code attached Use a GPT-5 layer to intelligently scan the transcript and identify 8 to 15 of the strongest, most shareable moments — returning a spec for each one including the start time, end time, and a suggested hook Generate platform-native captions for each clip in my brand voice — an Instagram Reel caption, a TikTok hook, a LinkedIn post, and an X thread Push all output into a Notion content calendar with status fields that are ready to action immediately Include a short Loom walkthrough video and a one-page runbook so I can operate the system myself without needing to contact you You are a strong fit for this role if: You naturally think in repeatable systems rather than one-time fixes You have already shipped n8n workflows that connect multiple external APIs with proper error handling and retry logic You understand what makes a short video clip actually land with an audience — not just where keywords appear in a transcript You are not the right fit if: Your only approach to video editing is manual cutting You cannot explain the logic behind how you would select clips from a transcript Bonus skills that will strengthen your application: Whisper and AssemblyAI transcription, ffmpeg for clip trimming, prompt engineering for brand voice replication, and experience with the Notion API. Deliverables: the working n8n workflow exported as a JSON file, a Loom walkthrough recording, and a one-page runbook in plain language.