I already have a first-cut codebase that an AI generated for a small desktop-web hybrid dashboard. It mixes Streamlit with a PySide/PyQt front end, runs on Python 3, and pulls Poppler and OpenCV in the background for PDF and image handling. What I need now is a developer who can step in, clean the code, and make the whole thing run exactly as intended. Core goal Turn the existing prototype into a smooth interactive dashboard that can visualise data coming from CSV files, live database connections, and a couple of light-weight REST APIs. The layout and widgets are sketched out; several functions compile but don’t yet talk to each other the way they should. Scope of work • Refactor the Streamlit and PySide/PyQt layers so they share state seamlessly (no duplicated logic). • Hook up the three data sources with reliable loading, caching where sensible, and error handling. • Implement interactive filtering, drill-downs, and simple export (CSV/PDF via Poppler). • Replace placeholder plots with real-time charts built in the library you think best fits (Plotly is fine, Matplotlib if lighter). • Integrate the existing OpenCV helper to display thumbnail previews when the data set contains image paths. • Package the project so it starts with one command — either `streamlit run app.py` or `python main.py` — and document the virtual-env setup in a concise README. Acceptance criteria – Sample CSV, a SQLite test database, and a public API endpoint all render side-by-side in a single dashboard page. – Switching data sources or filters triggers an update without breaking the UI. – No hard-coded paths; configuration lives in a single `.env` or config file. – The code passes `flake8` and mypy (basic level) and runs on Windows and macOS. Hand me the cleaned repo (Git preferred) plus short setup notes, and we’re done.