Engineering Index

Projects & System Case Studies

Production-certified desktop platforms, browser-native containerization engines, and generative AI pipelines built with high-availability, local-first principles.

Production / MS Store Certified2024 — Present•Lead Systems Architect & Desktop Developer

oaPOS — Windows Desktop POS & Real-Time Inventory Architecture

Microsoft Store ↗

A resilient, high-speed Windows desktop point-of-sale and inventory system built on a local HOST/CLIENT distributed architecture with zero cloud dependency and sub-10ms transaction latency.

Transaction Latency

< 10ms

Offline Resilience

100% Autonomous

Store Certification

Official MS Store

Terminal Sync

Real-time LAN Sockets

.NET 8WPF / DirectXC#PostgreSQLSQLite EmbeddedRedis CacheLocal Socket SyncMS Store

Deep-Dive Technical Case Study →

Active Production2024 — Present•Creator & Full-Stack Platform Engineer

Orvyna App Studio — Cloud Web2App Architecture & Native Store Delivery Engine

Live Platform ↗

An automated cloud compilation studio transforming responsive web applications into native-feel store-ready containers with direct Apple TestFlight / App Store Connect delivery and Google Play TWA AAB packaging.

iOS Pipeline

App Store Connect & TestFlight

Android Pipeline

Google Play TWA AAB

Bundle Generation

< 3 Seconds

Shell Footprint

< 150 KB

iOS TestFlightApp Store ConnectAndroid TWAWeb2AppTypeScriptVite

Deep-Dive Technical Case Study →

Interactive Tool / Active2024 — Present•AI Workflow Architect & Frontend Engineer

E-Commerce Image Enhancement Tool — AI-Powered Product Visual Studio

Launch Tool ↗

A generative AI pipeline and visual enhancement workbench that automates background replacement, lighting harmonization, and edge fidelity preservation for marketplace-ready product assets.

Production Turnaround

10x Faster

Edge Mask Accuracy

Sub-pixel Alpha

Lighting Coherence

Harmonized Shadows

Cost Reduction

-85% vs Studio

Generative AIDiffusion ModelsCanvas APIComputer VisionPrompt EngineeringGoogle Flow

Deep-Dive Technical Case Study →