Elena TCWEB SYSTEMS ENGINEER

Selected Engineering Project

Centralized Product
Information System

A centralized backend system that transforms supplier data into a single, reliable source of product information for e-commerce operations.

The system supports supplier imports, product normalization, pricing logic, AI-assisted SEO content generation, validation and controlled publication to an e-commerce storefront.

Project

Commercial product information system operating across multiple supplier data sources.

Role

Software architecture, backend development and system evolution.

Status

Active commercial project under continuous development.

Operational Context

Why the system was needed

The business depends on product information received from numerous independent supplier sources. Each source may use different identifiers, structures, update frequencies and delivery formats. As the number of suppliers and products grows, manual data management quickly becomes impractical.

Managing this directly inside the storefront would make the public website responsible for business logic, supplier decisions and data cleanup. The system was designed to move those responsibilities into a dedicated backend layer before information reaches the e-commerce platform.

Architecture Overview

High-Level System Flow

Supplier APIs / CSV / XLS

CSV
XLS

Import Layer

Collects data from multiple supplier sources in different formats and schedules.

Product Normalization

Converts supplier-specific data into a unified internal product model and identifiers.

Business Rules

Applies business logic, supplier priority, availability rules and data enrichment.

Pricing and Stock Logic

Calculates prices, availability and stock status in real time based on business rules.

AI-Assisted SEO Content

Generates and enhances product titles, descriptions and meta content with AI.

Validation

Validates data quality, content accuracy and business constraints before publication.

Controlled Publication

Publishes only validated and approved data to the storefront through a controlled workflow.

E-Commerce Storefront

Data is delivered to the e-commerce platform for customers.

Why It Was Designed This Way

Key Design Choices

A Single Internal Product Identity

Using supplier identifiers as the primary product key would tightly couple the entire system to external data sources. The same commercial product may exist under different identifiers, while a single supplier can also change identifiers over time.

Business Logic Outside the Storefront

Embedding business rules inside the storefront would make every future change dependent on the e-commerce platform. Moving business logic into the backend allows pricing, synchronization and publication workflows to evolve independently from the public website.

Independent Processing Pipelines

Supplier imports, pricing updates, content generation and publication are separated into independent processes. As the number of synchronization tasks grows, separating workflows prevents one operation from delaying the others and makes diagnostics significantly easier.

Controlled Publication

Automation is useful only while it remains predictable. Instead of publishing every generated change immediately, the system separates generation, validation and publication into distinct stages. This makes automation measurable and reduces the impact of unexpected results.

AI-Assisted Content Workflow

AI as part of the system, not as a standalone feature

Controlled generation

AI output is produced inside a backend workflow rather than directly inside the storefront.

Validation before use

Generated content is checked before publication. The system is designed so AI output does not automatically become product content without validation.

SEO-oriented content

The content workflow is designed for practical e-commerce use: product descriptions, metadata and controlled publication rather than generic text generation.

Human-controlled application

Automation reduces repetitive work, but final publication remains governed by system rules and controlled workflows.

Technology Stack

Implementation

Backend and data

PHP 8.4 · MySQL · custom backend logic

Integrations and processing

REST APIs · CSV · XLS · XML · Cron · LLM APIs

Confidentiality

Commercial Confidentiality

This overview intentionally omits client-specific business information, implementation details, internal metrics and proprietary logic. Its purpose is to demonstrate engineering approach while fully respecting commercial confidentiality. Additional technical discussion is possible only when it can be held without exposing confidential client information.

Contact

Let's discuss your project

Email LinkedIn
Get in touch