FALINDA

Platform overview

FALINDA is a live marketplace for consigning second-hand branded goods. Customers can discover products through a social-style experience on the web, iOS, and Android, while the operations team works through a dedicated back-office system.

The operational problem

Product information arrives through LINE as inconsistent free-form text, with varying specifications, spelling, categories, and missing attributes. Administrators previously had to interpret each message and enter the listing manually, limiting reported throughput to approximately 15 products per operator per hour.

AI-assisted intake workflow

1. LINE submissions enter an automation workflow built with n8n.
2. AI classifies the product, extracts attributes, and normalizes inconsistent text.
3. Validation rules identify incomplete or questionable information.
4. A human reviews the structured result before it enters the marketplace listing workflow.

Reported result

Listing throughput increased from approximately 15 to more than 200 products per operator per hour?a greater than 13-fold improvement without a corresponding increase in operator headcount.

The workflow shifted the primary operational constraint away from repetitive manual entry toward incoming product volume, validation, and market demand.

My contribution

• Developed and maintained the web, API, operations, and integration layers.
• Supported marketplace discovery, seller onboarding, moderation, reporting, and administrator approvals.
• Maintained the long-running Vue.js, Node.js, Express, and MongoDB platform.
• Designed and built the n8n and AI-assisted product-intake workflow with human review.

My verified contribution covers the web, API, and back-office systems; it does not claim authorship of the native iOS or Android clients.