Warehouse Order Platform
A two-sided B2B ordering system connecting a network of retail outlets with a central warehouse. Outlets browse a live catalog and build orders; the warehouse team manages inventory and drives every order through its lifecycle.
- Serving 100+ outlets with a large item catalog
- Full lifecycle: draft → processing → out-for-delivery → completed
- Bulk SKU import for large-scale catalog maintenance
- Role-based access for outlets vs. warehouse staff
Daily Closing Workflow System
An internal system that digitizes the retail money trail — from an outlet's daily takings to finalized finance records. Every sales figure, expense, and cash movement flows through a defined approval chain with full traceability.
- Structured Manager → Account → Finance → CEO approval flow
- Controlled status transitions to prevent data inconsistencies
- Dedicated audit trail — nothing changes silently
- Structured reporting visibility for financial tracking
POS Report Ingestion Pipeline
An ETL pipeline that pulls raw report files from the ARMS POS system over FTP, parses their proprietary fixed-format layout, and turns them into clean, queryable inventory-movement records — with import history and duplicate protection.
- FTP browsing, preview, and automated scheduled sync
- Custom parser for a proprietary report file format
- Hash-based duplicate detection on every import
- Encrypted credential storage — nothing in plaintext
SQL Voucher Tool
A zero-dependency, single-file web app that turns pasted spreadsheet rows into ready-to-run SQL for finance voucher generation — the entire application, logic, and storage in one self-contained HTML file that works offline.
- Paste-to-SQL with fully configurable column mappings
- Batch management: save, search, filter, reopen
- Multi-stream reporting and payout tracking
- No frameworks, no server — 100% offline-capable
Smart Motion Detection System
An IoT-based intruder-monitoring system for agricultural areas, combining low-cost hardware with on-device machine learning to detect and classify motion in real time — sending instant alerts to a farmer's phone.
- Built on Arduino Nano 33 IoT & ESP32 with PIR sensors
- TinyML model (Edge Impulse) classifying human / animal / no-motion at ~80% accuracy
- Real-time Wi-Fi & GSM alerts via the Blynk platform
- Hands-on hardware–software integration, testing & calibration
Want a closer look?
Source code for the business systems is kept private as they run live operations — but I'm happy to walk through the architecture, share screenshots, or give a live demo on request.
Request a walkthrough