Parts sourcing and RFQ platform
A sourcing workflow for an automotive parts catalogue: group parts into batches, choose supplier products for each quality grade, and generate a request-for-quote workbook for every supplier.
Anonymised professional work. Company names, internal systems, data and code are left out on purpose.
- My role
- Full-stack developer: the batch workflow in the front end, the API endpoints and the workbook generator.
- Type
- Professional
Overview
An internal tool buyers use to plan sourcing. Parts are grouped into batches, each part can carry several supplier products per quality grade, and the system produces a structured request for quote for each supplier.
The problem
Building requests for quotes by hand meant copying part details into a separate spreadsheet for every supplier, and the rules deciding which supplier product belonged to which grade were applied inconsistently.
Architecture
A React single-page app talks to an Express API backed by PostgreSQL. Server state is cached per batch with TanStack Query, and workbooks are generated on the server.
- Batch of parts
- Supplier products per grade
- Buyer review
- RFQ workbook per supplier
Approach
- Mapped how buyers actually work before designing screens, then laid the selection screen out as a grid of grades against categories.
- Moved the grade rules out of individual endpoints into one shared helper, so every screen and export applies the same logic.
- Shaped the data model so a part can carry several supplier products per grade instead of one.
Key features
- Batch overview, build sets and part selection screens
- Several supplier products per grade
- One formatted Excel workbook per supplier, ready to send
Challenges and decisions
The same business rule was implemented slightly differently in several places.
Consolidated it into a single shared helper used by every endpoint, which made results consistent and the rule easy to test.