AI-Powered Football Video Editing Agent (Claude Code/Gemini/ChatGPT Codex)

I am looking for an experienced AI engineer to build an AI video editing agent that I can run using Claude Code or OpenAI Codex. The deliverable should include all the agent prompts, workflow, and code so that I can run and maintain it myself. The AI agent’s primary purpose is to automatically generate football match highlight videos from full-length match recordings with minimal manual intervention. BACKGROUND Currently, my recordings include a scoreboard overlay containing information such as the match timer and score. Example video: 
https://youtu.be/TKF48j3BNBo?is=9uD5gftj2STL8oV3 In the future, I may also record matches in their raw format without any overlays. The solution should therefore be designed to support both scenarios. VIDEO PROCESSING REQUIREMENTS The solution should ideally support the following workflow: ● Read videos directly from a YouTube link, or ● Accept videos uploaded directly from my iPhone. ● Handle full match recordings of up to 80 minutes. ● If LLM upload limits are a constraint, the agent may run remotely on a PC or server while I control and initiate the workflow from my iPhone. ● Once processing is complete, the generated highlights should be automatically uploaded to YouTube (or be ready for upload). My preference is to manage the entire workflow from my iPhone without needing access to a laptop while travelling. HIGHLIGHTS GENERATION The AI should automatically identify and produce highlights containing: ● Goals ● Near misses ● Shots on target ● Saves ● Other exciting attacking moments Goal Replay For every goal: ● Include the original clip. ● Immediately follow it with a replay. ● The replay should: ● be slowed down, ● zoom into the scoring action, ● focus on the important area of play. Highlight Overlays During highlight clips, display contextual overlays such as: ● GOAL ● SHOT ● SAVE ● CLOSE CHANCE ● Other relevant event labels where appropriate. MILESTONES Milestone 1 – Videos with Overlays Develop an AI workflow that processes football match videos containing existing score and timer overlays to: * Detect goals and other exciting moments, including shots on target, saves, near misses, and key attacking plays. * Automatically generate a highlight reel containing only the detected key moments. * For every goal, include the original goal clip followed immediately by a duplicated replay in slow motion with a digital zoom focused on the goal-scoring action. * Preserve the existing match overlays (e.g. score and timer) from the source video throughout the highlights. * Add additional event overlays during the highlights (e.g. GOAL, SHOT, SAVE, CLOSE CHANCE) where appropriate. * Export the completed highlight video in a format suitable for direct upload to YouTube. Milestone 2 – Videos without Overlays Extend the solution to process raw videos without any overlays. The AI should: ● Detect goals and exciting moments using computer vision and/or audio analysis. ● Generate the same style of highlight video as Milestone 1. ● Automatically create event overlays (e.g. GOAL, SHOT, SAVE, CLOSE CHANCE) since they are not present in the original footage. DELIVERABLES The successful freelancer should provide: ● A complete AI agent that runs on Claude Code or OpenAI Codex. ● All prompts, code, configuration files, and documentation. ● Instructions for setup and execution. ● A workflow that is easy for me to operate and maintain. ● Well-structured, reusable code so I can continue enhancing the solution in the future. Experience with AI agents, computer vision, LLM workflows, and automated video editing is highly preferred. REFERENCES Claude Viewing Video Skills https://youtu.be/QZMljuD10sU?is=KgNbWpG0iWxCsp4B