FastAPI + React + ML Starter Template

[roohan-514]

Spent the weekend building a full-stack ML starter template so I never have to do it from scratch again.

Every time I start a new ML project, I end up doing the same dance — setting up FastAPI, wiring up a React frontend, writing a Dockerfile, configuring MLflow, setting up CI. It's maybe an hour of boilerplate, but it adds up across projects. And honestly? It's boring.

So I finally sat down and extracted it into a proper template.

What's inside:

• FastAPI backend with a working scikit-learn Iris classifier (train + predict endpoints)
• React + Vite + TypeScript + Tailwind frontend with a form that calls the predict API
• Docker Compose that spins up backend, frontend, and MLflow in one command
• GitHub Actions CI that validates both backend and frontend on push
• Production-ready nginx config for the frontend container

The idea is simple: clone it, swap out the model, change the frontend form fields, and you're live.

https://github.com/roohan-514/fastapi-react-ml-starter