Hey Dev Community 👋
We’re two brothers building ModelShip — an AI-powered platform to help developers and ML teams go from raw, messy data to labeled, export-ready datasets using smart, human-in-the-loop workflows.
We’re building publicly, not open-source. And we’re sharing every commit, every test, and every challenge as we go, so you can watch us win, or even help us get there faster.
🧠 Why We're Building This
If you've ever tried training an AI model, you know:
The worst part is prepping the data.
Labeling is:
Tedious
Error-prone
Slow
And too often, done in spreadsheets 🙃
We're building ModelShip to fix that.
🧱 What ModelShip Will Do (MVP)
✅ Upload your dataset (CSV or JSON)
✅ Auto-label rows using GPT-based AI
✅ Let you manually review or override labels
✅ Export labeled data to CSV or JSON
✅ Train custom models on your own data
This isn’t just a backend.
This is a real product, with a real vision:
Label smarter, train faster, ship AI products quicker.
🧑💻 What We've Built So Far
We're building the whole stack:
FastAPI (Python) for services
MongoDB Atlas for data
Cloudinary for file upload
OpenAI (GPT-3.5) for smart labeling
Frontend (Next.js)
We’re pushing backend services live now to GitHub — not as open source, but as public proof of progress.
🛠️ Check it out here:
🔗 https://github.com/ModelShip25/ModelShip
💬 Why We’re Sharing This
We want real feedback from real devs
We’re looking for volunteers or contributors who want to build cool stuff
We believe in showing our work, not just launching with a logo
🤝 How You Can Help
Watch or star the GitHub repo
Follow us on Twitter/X: [x.com/Teeshine18]
Comment here with your thoughts or ideas
DM us if you want to be part of the team (even part-time)
We’re open to:
Devs who want to code
Testers who want to break things
Designers with strong product sense
Friends who believe in building something from scratch
This isn’t another SaaS wrapper.
This is the infrastructure for better AI labeling — built by devs, for devs.
And we’re doing it with anyone who wants in.
Top comments (0)