Offline-First Fisheries Ledger & Voice AI Assistant Development by Sahibul Nuzul FirdausOffline-First Fisheries Ledger & Voice AI Assistant Development by Sahibul Nuzul Firdaus

Offline-First Fisheries Ledger & Voice AI Assistant Development

Sahibul Nuzul Firdaus

Sahibul Nuzul Firdaus

Offline-First Fisheries Ledger & Voice AI Assistant

Role: Lead Mobile & AI Engineer

Tools & Skills:

Flutter, Android, Supabase, PostgreSQL, Row-Level Security (RLS), PowerSync / Offline Data, Gemini Multimodal API, Speech-to-Text (STT)

Project Overview:

Built an offline-first mobile operations and financial ledger tailored for vessel operators and maritime businesses. The application enables crews to log operational expenses (fuel, ice, provisions, catch yields) at sea under zero-connectivity conditions, featuring real-time receipt scanning and a voice-driven AI assistant.

The Challenge / Problem:

Maritime vessel managers operate in remote offshore areas with little to no cellular connectivity, causing critical transaction logs and catch records to be lost or delayed.
Manual bookkeeping resulted in prolonged revenue calculations, calculation errors during profit-sharing settlements (bagi hasil), and slow record reconciliation.
Entering detailed multi-currency and unit data on small mobile screens while on active duty was slow and error-prone.

The Approach & Architecture:

Offline-First Synchronization: Engineered a robust local-first database replication architecture using Supabase and PowerSync/SQLite, allowing seamless offline CRUD operations with conflict-free cloud reconciliation once connectivity returns.
Multimodal Voice & Receipt AI: Integrated Google Gemini multimodal APIs and on-device STT pipelines to enable hands-free voice logging and automated receipt parsing directly into structured transaction drafts.
Complex Settlement Logic: Built an automated net-profit and multi-party share distribution calculation engine with instant formatted report generation ready for WhatsApp dispatch.
Hardened Security & Isolation: Implemented granular PostgreSQL Row-Level Security (RLS) policies preventing cross-vessel and cross-tenant data leaks.

Key Results & Impact:

Successfully deployed an operational pilot running smoothly under 100% offline conditions at sea.
Accelerated voyage financial settlement and profit-sharing calculations from hours of manual tabulation to under 30 seconds.
Achieved zero data loss across hundreds of offline-logged transactions during real-world maritime pilot runs.
Like this project

Posted Aug 28, 2026

Developed an offline-first ledger app for maritime operations, improving transaction logging and settlement speed.