Observe
Find where runtime behavior and team expectations diverge.
Backend & Platform Engineer
Backend and platform engineer connecting event-driven research, production delivery, and home-lab operations.
Workbench
Composite Contract Testing Service — my NTOU thesis system and APSEC 2022 paper implementation. It combines Pact contracts, event-log sequence validation, and state-model checks to catch missing transitions and execution failures in event-driven microservices.
APSEC 2022 · TCSE Best Chinese Paper · 8k+ LOC Java · Spring Boot · Docker · Fly.ioAnonymous Discord reply bot built for the NTOU CS Discord context. It grew from a course project into a containerized service with CI/CD, Docker Hub image flow, and Fly.io deployment.
Full lifecycle side project · 10 stars Docker · Linux · NAS · MariaDB · SDNHome-lab service configuration and operations notes. The broader environment runs 30+ services across storage, DNS, monitoring, media, albums, and home automation with VM/container integration.
30+ services · real operating environment Shell · Docker · IPMI · ESXi · NVIDIADell PowerEdge fan control through iDRAC/IPMI with Docker deployment, NVMe/GPU temperature inputs, configurable thresholds, and a recent README/code audit around fail-safe behavior.
Dell R730XD tested · 10 starsBuild pipeline
Find where runtime behavior and team expectations diverge.
Turn flows into contracts, event paths, deployment notes, or testable boundaries.
Ship the smallest durable system that reduces operational uncertainty.
Document, monitor, and remove friction before complexity compounds.
Material history
Ming-Chyuan MIS work creates a long-running baseline: software has to survive legacy systems, manufacturing workflows, and maintenance pressure.
B.S. and M.S. work converge on service-oriented architecture, event-driven testing, and maintainability under research constraints.
CCTS turns consumer-driven contracts, event logs, and state models into a composite test service for event-driven microservices.
Personal-data authorization and valuation/payment platform work spans logging, payment, blockchain, CI/CD, fintech integration, and AKS delivery.
Self-hosted systems, server automation, and public PRs keep the engineering habit grounded in operational feedback.
Shop notes
Profile, projects, metrics, insights, styles, search, feed, OpenAPI.
all variantsKeyboard-accessible search over headings, insights, projects, and APIs.
all variantsClient fetches public GitHub profile and falls back gracefully.
progressive enhancementImpact bars, timeline, article taxonomy, and role-fit matrices.
all variantsDesign rationale, technical notes, Mermaid diagrams, branch/deployment instructions.
ten variantsCCTS matters because it connects contract definitions with event-log paths and state-model assertions, closing the gap between interface agreement and actual behavior.
6 min readThe PDAS work shows why deployment, CI/CD notes, infrastructure constraints, and customer handoff are not afterthoughts; they decide whether the system can be trusted.
5 min readPortfolio intelligence layer
This shared layer turns every variant into a richer portfolio surface: API endpoints, command search, live GitHub enrichment, article cards, timeline context, and machine-readable data.
Profile, projects, metrics, insights, styles, feed, search, and OpenAPI are generated at build time.
The client can fetch current public profile data without requiring secrets. If the network fails, static evidence remains complete.
Ming-Chyuan MIS work creates a long-running baseline: software has to survive legacy systems, manufacturing workflows, and maintenance pressure.
B.S. and M.S. work converge on service-oriented architecture, event-driven testing, and maintainability under research constraints.
CCTS turns consumer-driven contracts, event logs, and state models into a composite test service for event-driven microservices.
Personal-data authorization and valuation/payment platform work spans logging, payment, blockchain, CI/CD, fintech integration, and AKS delivery.
Self-hosted systems, server automation, and public PRs keep the engineering habit grounded in operational feedback.
Engineering journal
CCTS matters because it connects contract definitions with event-log paths and state-model assertions, closing the gap between interface agreement and actual behavior.
The PDAS work shows why deployment, CI/CD notes, infrastructure constraints, and customer handoff are not afterthoughts; they decide whether the system can be trusted.
Running storage, DNS, monitoring, media, albums, and automation changes how architecture feels: failures become personal, and clarity becomes operational leverage.
Integration, localization, and documentation fixes reveal a practical collaboration style: learn the boundary, reduce friction, and leave a clearer path behind.