Full-Stack SaaS Developer Needed — AI-Powered Twitch Clip Platform I'm building EchoClips, a SaaS platform that uses AI to automatically detect and clip the best moments from Twitch VODs. Think ClipFinder.org but with a real AI pipeline behind it — VideoMAE + CLIP models running on a dedicated GPU. This is a complete, well-scoped project with a full technical briefing attached. Please read it before applying. --- WHAT YOU'RE BUILDING • Landing page (dark theme, purple accents, hero, pricing, testimonials) • Timeline editor — scrubber, waveform, draggable start/end, keyboard shortcuts • AI clip detection — VideoMAE + CLIP models suggest top 3–5 moments per VOD • Auth — Email/password, Google OAuth, Twitch OAuth • Credits & subscription system — $4.99 one-time / $12.99 monthly / $9.99 annual • PayPal integration (subscriptions + one-time orders) — Stripe is not available in Morocco • Referral system — custom links, configurable % discount for referred users, cash earnings for referrers, referrer dashboard, withdrawal requests with admin-controlled limits • Admin dashboard — user management, VIP/referrer roles, revenue stats, AI job monitor, pricing controls, testimonials CRUD, chatbot config • Help chatbot — embedded support bot • Clip export — MP4, 9:16 TikTok, 16:9 YouTube --- REQUIRED TECH STACK Frontend: Next.js + React + Tailwind CSS Backend: Laravel, Django, or FastAPI (your choice) Database: Supabase (PostgreSQL + real-time) AI/GPU: Python + Hugging Face (VideoMAE, CLIP, Whisper) on RTX 3080 Queue: Redis + BullMQ (priority by tier) Storage: Cloudflare R2 Video: FFmpeg + YOLO + MediaPipe Payments: PayPal REST API Monitoring: Prometheus + Grafana --- WHO I'M LOOKING FOR • You have shipped at least one full SaaS product end-to-end • You are comfortable with GPU model deployment (Hugging Face, Docker) • You've worked with BullMQ or similar job queues • You write clean, maintainable code — not just code that works • You communicate clearly and flag blockers early Please do NOT apply if you haven't read the attached briefing or if VideoMAE/CLIP deployment is outside your experience. HOW TO APPLY To filter out copy-paste proposals, start your message with the word ECHOCLIPS and answer these two questions: 1. Have you deployed Hugging Face models on a GPU server before? Describe the project briefly. 2. What backend framework would you choose for this project and why? Proposals that don't follow this format will not be reviewed. --- Full technical briefing is attached. Budget is negotiable based on your experience and approach. Looking to start within the week.