A synthetic inventory reconciliation demo built with AI coding assistance. It shows how I approach a scoped data-cleanup task: explicit matching rules, preserved inputs, and an audit of every decision.
The Python script matches warehouse plus text SKU, preserving leading zeros. It updates only unambiguous, valid quantities and flags duplicates, missing matches and invalid values. It preserves destination order and the original files.
The example produces 3 updates, 1 unchanged match and 6 target rows for review, plus a source-only audit record. The repository contains the implementation, invented input data, output previews and 10 automated tests, including a 50,000-record combined scale case.
This is a demonstration, not a completed client assignment. The workbook is a static preview; this sample does not connect to live Excel or SharePoint.
Source, tests and results: https://www.twine.net/signin
Portfolio: https://www.twine.net/signin
Contribution: AI-assisted Python implementation, test coverage, synthetic examples and handover documentation prepared for my scoped reconciliation service.
- Full Stack Developer