Senior AI Video Systems Engineer-Automated Sports Highlight & Mixtape Platform

Senior AI Video Systems Engineer — Automated Sports Highlight & Mixtape Platform About the Project We are building an AI-powered sports technology platform focused on youth and amateur sports. We are looking for an experienced Senior AI Video Systems Engineer / Programmatic Video Engineer to lead development of an automated sports highlight and mixtape generation system. This is not a traditional video-editing role and is not simply an LLM integration project. The system will ingest sports video clips, structured game/event metadata, player information, audio assets, and editorial instructions and automatically produce professional-quality short-form sports videos. Think: Sports highlights + intelligent clip selection + music synchronization + graphics + automated editing + cloud rendering. The ideal engineer has experience building production media pipelines with technologies such as Shotstack, Remotion, FFmpeg, TypeScript/Node.js, cloud infrastructure, and AI APIs. Experience with sports video, computer vision, automated content creation, or media rendering systems is highly desirable. What You Will Build You will be responsible for developing a production-ready video generation pipeline capable of taking structured instructions and source media and turning them into finished sports mixtapes automatically. The system should eventually support multiple editorial styles, including: - High-energy sports hype videos - Player recruiting highlights - Player keepsake/season videos - Social-media highlight packages - Team highlight videos - Sponsor-supported branded content Videos will generally be short-form content optimized for approximately 30–90 seconds, although the underlying architecture should support additional formats. The system must be capable of operating automatically at scale rather than requiring a human video editor for every output. Core Responsibilities Programmatic Video Rendering Build a reusable video rendering architecture using technologies such as: - Remotion - FFmpeg / FFprobe -Shotstack/Shotstack API - TypeScript / JavaScript - Node.js - Cloud-based rendering infrastructure Develop reusable components for: - Video clips - Transitions - Player graphics - Score graphics - Statistics - Titles - Intros/outros - Sponsor graphics - Calls to action - Logos and branding - Dynamic text - Vertical, square, and landscape formats The system should assemble videos from structured data rather than requiring manually coded compositions for every video. The engineer should be comfortable evaluating when to use Shotstack’s JSON-based editing/rendering API, reusable Remotion compositions, direct FFmpeg processing, or a combination of these approaches based on rendering flexibility, cost, throughput, reliability, and maintainability. Intelligent Video Assembly Develop logic for: - Clip selection - Clip ordering - Clip duration - Intelligent trimming - Highlight prioritization - Timeline construction - Pacing - Replays - Slow motion - Speed ramps where appropriate - Dynamic transitions - Story sequencing Structured metadata may include information such as: - Player - Team - Game - Score - Event type - Timestamp - Play importance - Statistics - Clip quality - Editorial priority The architecture should be capable of using this information to make intelligent editing decisions. Music & Audio Synchronization Develop an audio-processing pipeline capable of: - Audio analysis - BPM/beat detection - Beat-grid generation - Music-aware clip timing - Cutting video on beats - Synchronizing transitions to music - Audio ducking - Sound effects - Mixing - Drops and emphasis points - Other sports-mixtape-style audio treatments Experience building beat-aware automated editing systems is particularly valuable. AI Integration Integrate modern AI models/APIs into the editorial workflow. Potential responsibilities include: - Structured-output generation - Editorial planning - Clip ranking - Story sequencing - Metadata analysis - Caption generation - Video composition instructions - Quality-control assistance Experience with APIs from platforms such as: - Anthropic Claude - Google Gemini - OpenAI - Other multimodal or video intelligence systems is useful. We are specifically interested in engineers who understand how to use AI for decision-making and orchestration while maintaining deterministic production pipelines. Rendering Infrastructure Build a scalable rendering system including: - Render queues - Worker architecture - Job management - Retry logic - Idempotent processing - Error handling - Progress tracking - Webhooks - Render manifests - Logging - Performance monitoring - Cost monitoring The system should eventually be capable of processing a significant number of video-generation jobs without manual intervention. Media Pipeline Work with: - Cloud object storage - CDN delivery - Video transcoding - Codec management - Video normalization - Resolution management - Aspect-ratio conversion - Thumbnail generation - Media metadata extraction Experience with AWS or another major cloud provider is strongly preferred. Required Experience We are looking for a senior-level engineer, not someone learning these technologies during the project. Strong candidates should have meaningful experience with several of the following: - Remotion or another programmatic video framework - FFmpeg / FFprobe - TypeScript / JavaScript - Node.js - React - REST APIs - Webhooks - JSON-based workflow specifications - Cloud rendering - AWS, GCP, or Azure - Object storage - Queue/worker architectures - Media processing - Video transcoding - Audio processing - AI/LLM APIs - Production backend systems You should be comfortable owning a complex technical system rather than waiting for step-by-step implementation instructions. Strongly Preferred Experience with any of the following will make you particularly interesting to us: - Automated video editing - Sports highlight generation - Sports technology - Sports video analysis - Music synchronization - Beat detection - Computer vision - OpenCV - Python - Player tracking - Action recognition - Video segmentation - Multimodal AI - Automated social-media content generation - Large-scale rendering pipelines Computer vision experience is valuable, but you do not need to be a computer-vision researcher to qualify for this role. Architecture Philosophy We are not looking for a system where an AI model generates arbitrary video-editing code for every request. We want a reliable production architecture consisting of: Structured editorial instructions → deterministic media pipeline → reusable video components → cloud rendering → automated QA → finished media AI should enhance editorial intelligence while the underlying rendering system remains reliable, observable, and testable. If this architecture makes sense to you, you are likely the type of engineer we want to speak with. Initial Engagement We expect the initial engagement to be approximately 30-60 days, with the potential for substantial ongoing work. This could become a long-term technical ownership opportunity for the right person. The initial project will focus on: 1. Building the core rendering engine 2. Converting structured editorial instructions into finished video 3. Creating reusable sports-video components 4. Implementing multiple editorial styles/templates 5. Building music synchronization 6. Establishing cloud render infrastructure 7. Implementing monitoring, error handling, and automated QA 8. Preparing the system for integration into our larger sports platform You will work alongside our existing development team, but you will have significant ownership of the media-generation architecture. When Applying Please begin your proposal with the words: SPORTS MIXTAPE This allows us to identify applicants who have actually read the posting. In your proposal, please answer the following: 1. Have you built a programmatic or automated video-generation system before? Describe it. 2. What experience do you have with Remotion and FFmpeg? 3. Have you built systems that synchronize video edits or transitions with music? 4. What would your high-level architecture be for automatically generating a 60-second sports highlight video from 10–20 source clips plus structured metadata? 5. What technologies would you use for the rendering queue and worker infrastructure? 6. What experience do you have integrating Claude, Gemini, OpenAI, or other AI models into production applications? 7. Have you worked with sports video, computer vision, or automated highlight detection? 8. Please provide links to actual video-generation systems, rendering platforms, automated media tools, or relevant code/projects you have personally worked on. 9. Clearly identify which portions of those projects you personally designed or implemented. 10. What is your availability over the next 60–90 days? Please do not apply if your primary experience is manual video editing, basic AI wrappers, prompt engineering, or generic frontend development. We are looking for someone who can architect and build the underlying automated media system. The fixed cost stated is just to satisfy the posting. This job will be paid on a negotiated hourly rate.