Zin Min
Senior Software Engineer
6+ years building and operating production systems: event-driven platforms, data pipelines, and AI-integrated products. The work tends to run end-to-end, from architecture through deployment and operations.
Currently Founding Engineer at NOSBAAN, an Airbnb Superhost-recognised monthly rentals business in Bangkok, where I build and run the whole platform. Alongside that, I build MonOCR, an OCR system for Mon, a UNESCO-classified vulnerable language with almost no prior dataset, model, or toolchain. I built it from collecting the data through to training the models.
Selected Work
MonOCR — OCR for the Mon Language
Built an OCR system for Mon, a UNESCO-classified vulnerable language. With no existing datasets, OCR models, or supporting toolchain available, the project began with collecting and validating training data before building the models and deployment infrastructure from the ground up.
The system includes a compact mobile model (~11.4M parameters) and a larger server-side model (~54M parameters), enabling deployment across both on-device and backend environments. Supports ONNX and CoreML export, with a continuous feedback loop that collects validated samples to improve recognition accuracy over time.
Data Platforms
Built a public civic data platform serving ~10K active users (2.8K peak daily), and a separate research platform serving near-real-time policy and research data.
Replaced an Airflow-based orchestration layer with a lightweight custom ETL pipeline, reducing data refresh times from 1–2 hours to under 5 minutes while simplifying operations and lowering maintenance overhead.
System Migration & Cost Reduction
Led the migration of a legacy platform to a simpler Next.js + Baserow architecture, reducing infrastructure costs by 33%.
Rebuilt the data pipelines and added backup and recovery, which the previous setup lacked, improving reliability and reducing operational risk.
Event-driven ERP & AI Messaging Platform
Designed and built the core ERP platform, public portal, and AI-assisted messaging system that powers daily operations at NOSBAAN.
The platform uses an event-driven architecture with asynchronous processing and RAG-based AI workflows to isolate failures, handle resource-intensive workloads, and improve system resilience. Even on modest infrastructure, it sustains 50+ req/s at P95 < 1.5s under load.