Amirhossein Ghanipour Amirhandeh

Sole engineer across the full product surface for a 3-person consumer fintech startup: ML inference services, Flutter client, Supabase backend, marketing site, and supporting AWS/GCP infrastructure. Distilled a recommendation model to run entirely on-device by implementing a custom character-level inference runtime (avoiding TFLite/ONNX overhead) and ensuring personalisation data never leaves the handset. Designed and operated a receipt-scanning inference API, rebuilt releases as an automated reproducible pipeline across both clouds, and introduced monitoring/alerting for inference failures.

Owned the Flutter application for iOS and Android (published on App Store and Google Play). Implemented clean architecture using Riverpod and added native platform code when the Flutter ecosystem was insufficient. Built the receipt-scanning computer-vision pipeline to extract line items from photographed receipts and implemented an offline-first Supabase data layer with conflict resolution. Delivered the app in 10 languages with full RTL mirroring for Persian and Arabic, profiled/optimized startup and render performance, and automated store releases via Codemagic.

Built custom CRM, Sales, and Accounting modules on the Odoo framework using Python and JavaScript/XML. Adapted a general-purpose ERP to client-specific workflows.