Tech Stack
- language: SwiftUI, UIKit
- database: Core Data
- AI: Open AI Whisper
- Production Monitoring (crashlytics): Sentry
- Addson: Ifrit, Swiftlint, Speech
Description
Managing food inventory effectively is a common challenge for small business owners, especially for food stall (warteg) operators. Traditional methods often result in food wastage, inaccurate stock tracking, or a lack of real-time insights.
My team set out to design and develop a mobile application, UStock, aimed at simplifying food inventory management. The goal was to empower warteg owners with an easy-to-use platform to track stock levels, minimize waste, and streamline operations, especially users with limited technological experience.
As a developer on the team, i was responsible for the following action:
• Designed and developed UStock using SwiftUI, creating an intuitive UI for effortless navigation and accessibility.
• Implemented gesture-based controls for fast stock updates, making it easy for users to adjust quantities with a few taps.
• Integrated a voice command feature using Apple’s Speech Framework and a quantized Whisper model, enabling users to update inventory hands-free with multilingual support (Indonesian and English).
• Built a dynamic notification system that sends daily alerts to remind users to restock low inventory items.
• Leveraged Core Data to enable detailed history tracking and reporting, allowing users to review past inventory trends for better decision-making.
• Incorporated error and bug tracking using Sentry, optimizing app performance and ensuring reliability in production.
I also doing optimization and accessibility on the app to make it more efficient and user friendly:
• Reduced app file size by 50% through code optimization and quantitazion on whisper models.
• Reduced memory usage up to 90% by using speech SDK when the user is online and whisper model when the user is offline.
• Boosted voice command accuracy by implementing fuzzy search (ifrit) with a 60% similarity threshold.
• Enhanced accessibility with Dynamic Text feature, ensuring a user-friendly experience for older adults.
• Successfully published the app on the App Store with 2 localization (English and Indonesian), increasing accessibility for a broader audience.
Assets
