
Full-Stack Developer to Build & Package End-to-End “Data Copilot” Web App (CSV/Excel → Insights)
Upwork
Remote
•10 hours ago
•No application
About
Seeking an experienced developer to transform an non functional / buggy MVP into a production-ready product. The project involves enhancing a Streamlit + DuckDB + Pandas prototype to include file uploads, schema preview, natural language and SQL querying, visualizations, export options, containerization, deployment readiness, authentication, branding, scalability, and documentation. What the app should do File Uploads: Users can upload large CSV/Excel files (~200MB). Schema & Preview: Show schema + sample rows in a clean UI. NL & SQL Querying: Users type questions in plain English (NL → SQL translator). Or directly run SQL queries (DuckDB backend). Visualizations: Return results in both table + chart (bar, line, pie). Export: Option to download results as CSV/Excel. Beyond the MVP (packaging/productization) Containerization: Package the app in Docker/Kubernetes (not just Streamlit Cloud). Deployment-ready: Prepare for private cloud/on-prem or SaaS hosting. Authentication & Multi-user: Add login + session isolation. Branding/UI polish: Add a simple, modern design (my brand colors/logo). Scalability: Ensure efficient handling of large datasets. Documentation: Clear README and setup guide for me/future developers. Tech stack (open to discussion) Python, Streamlit (or alternative modern frontend if better). DuckDB (for fast local SQL queries). Pandas / Polars for data handling. Plotly/Matplotlib for charts. Docker for packaging. Optional: LangChain or similar for better NL → SQL translation.