Convert Notebooks Into Interactive Webapp
Publicada el 2026-07-22
Descripción de la oferta
I have a collection of demonstration notebooks here: https://github.com/DHANUSHGCODE/data-science-docs-examples. Your task is to turn each of those notebooks into a production-ready, containerised web application. Backend • Re-factor the code that currently lives in the notebooks into clean Python modules and expose the endpoints through FastAPI. All model inference, preprocessing and analytics logic should be callable both from the UI and via a REST route. • Write concise docstrings and ensure unit tests are in place so the service can slot into a CI pipeline later. Frontend The public face of the app must be built with Streamlit and deliver three key capabilities: • Interactive model visualisations (plots, metrics, confidence intervals, etc.) • Dataset upload and download so users can bring their own CSV/JSON and retrieve processed outputs. • Real-time analytics that update live as a user tweaks parameters or streams new data. Packaging & DevOps Package everything into a Docker image that runs with one command. A simple Docker-Compose file is enough for local orchestration. Please keep Git history clean and open a pull request for each logical chunk of work so I can review incrementally. Acceptance criteria • Each notebook is represented by at least one FastAPI route and a matching Streamlit page. • All three frontend features above are functional and tested. • The container starts, serves the Streamlit UI at / and the API at /api, and passes all included tests with `docker compose up`. If this sounds straightforward to you and you have solid MLOps experience, I’m eager to review your approach and timeline.
Skills
Fuente original: freelancer