shipped a production-grade AI video dubbing studio. it’s fully built, live, and ready for scale—now looking for a co-founder / growth partner to help run it.
the backend is an absolute beast .built a custom distributed MLops architecture running on fastapi and Modal to dodge paper-thin API wrappers:
• frontend/storage: Next.js pushes to Cloudflare R2 via presigned URLs, backed by Supabase Storage and SQL-based chains for handling state and assets.
• worker pool: CPU containers extract audio via FFmpeg, while heavy L4 GPUs run isolated Demucs + WhisperX stacks.
• voice cloning: Fully self-hosted Zonos-2 pipelines utilizing VRAM snapshots to instantly bypass container cold starts.
• data transit: A shared storage network volume shares files across containers on the fly before final FFmpeg assembly.
the core engine and infrastructure are 100% completed.
what i need('specially'): i m a solo technical founder driving the infra. looking for someone to own the product side—managing users, building marketing pitches, and driving growth.
if u fully understand the tech and wanna help scale this into a real user base, DM me for a 7-day beta token to test the pipeline and talk collaboration.
modal script : github.com/williamjosh-dev/AI-Dubbing-Tool/blob/…/modal_app.py
deploy at vercel : ai-dubbing-tool.vercel.app ("note : u cant upload a video to my backend cuz i impelled a usage based beta token so u need to dm first before testing")
also have n a tweet at #C01GPEL6CGM: hackclub.slack.com/archives/C01GPEL6CGM/p1791791493541169🚗❄️giving LLMs a "visual Brain" via a 4d Physics Sandbox
llms struggle with spatial geometry because they process text tokens, not dimensions. I am building a physical sandbox framework to act as an external visual cognitive engine for AI.
how the loop works:
• Code Gen: LLM generates logic to spin up a 4D cloud workspace (3D + time).
• Offloading: Heavy orbital mechanics and structural math offload to my custom GPGPU physics engine.
• Self-Correction: Live telemetry logs from simulation failures pipe back to the LLM to auto-patch its code.
Update: Prototype Done, Moving to Realism!
The agentic loop prototype works. Now, I am building the first 4 planets of the solar system using:
• NASA Maps: Real elevation data in a Three.js frontend.
• Atmospheric Clouds: Custom GLSL shaders for dynamic layers.
• Newtonian Orbitals: A multi-body backend gravity matrix.
Pushed the core builds to GitHub: github.com/williamjosh-dev/Physical-Sandbox-v1.0-
I'm 15 and currently trying to balance massive planetary scales without breaking my WebGL views. Let me know what you think ...
🚗
❄️
