Textile Production ERP
The problem
Lack of real-time traceability on the textile production line, causing undetected bottlenecks and outdated inventories.
The solution
Comprehensive management system (ERP) focused on the production section, allowing shop floor control, material traceability, and station optimization.
Architecture
React application and PostgreSQL. The floor records stations, materials, and progress. The schema privileges traceability over a generic CRM. The management dashboard reads the same store; there is no second system “for reports.”
How it was measured
Figures reported by the ERP engineer at Towel S.A. de C.V., Oct 2025–present. The 40% compares workflow visibility (stations and materials visible in the system versus boards and sheets) before and after deployment. Paper elimination is plant-floor observation, not a published sheet count. No third-party audit. No screenshots on this page: the system runs on the floor and was not photographed for the portfolio.
What does the textile floor see in real time that used to be paper?
Stations, materials, and order progress in a React/PostgreSQL ERP. The bottleneck no longer lives on a sheet someone updates at the end of the shift. Management queries the same record from the dashboard.
The work
The ERP is the current job. Towel S.A. de C.V., from October 2025. The plant did not have real-time traceability: the bottleneck showed up after it had already happened, inventory in the system did not match the floor, and the recording medium was paper. A generic “whole business” ERP was not the request. The request was the production section: stations, materials, progress. Shift by shift, not a management dashboard that updates overnight.
I owned architecture and implementation in React and PostgreSQL. The floor loads state. The schema is built to follow a lot, not to be a customer CRM. The business-intelligence dashboard, on another page of this site, reads this same store. If the two faces diverged, management would ask for a spreadsheet again and the floor would go back to paper. That is the only reason the internal link between the two URLs exists.
The 40% visibility figure is the comparison I made as operator between “knowing which station is blocked now” and “hearing about it in the afternoon meeting.” It is not a consultant KPI. Paper nearly disappeared because the record lives in the system; I did not count reams. A recruiter who wants visual proof will not find it here: I did not invent screenshots. The system runs in a plant, with people and machines, and there is no public staging environment to photograph without permission. If real photos exist later, they get added. Until then the prose carries the weight.
If this is the first page a hiring manager opens, that is fine. It is the current employment, the store shared with the dashboard, and an impact that states its method and the absence of screenshots in the same breath. The rest of the portfolio can wait.