NTCU CS Department Website

Team project — I served as project lead for NTCU’s Computer Science department website.

NTCU CS Department Website is the official site of the Department of Computer Science at National Taichung University of Education (NTCU) — the department’s public face for faculty, curriculum, announcements, and resources. I was honored to be trusted as the project lead during my undergraduate years, building and running the site with a team.

My role

As lead, I was responsible for:

• Coordinating the team — assigning tasks, reviewing members’ code, and approving merge requests.
• Evolving the site to meet the department’s changing requirements.
• Operating the servers — keeping the deployment machines running.
• Mentoring junior members — passing on the development and operations know-how so the site would always have maintainers, making handover and long-term continuity possible.

Architecture — built for maintainability

Because a departmental site has to run for years, the hardest and most important part was getting the architecture right. With mentorship from a senior who guided me through it, we designed the system around maintainability and extensibility, so day-to-day operation stays simple enough for anyone to run:

1. I assign the work; a teammate implements the change.
2. I review the code and approve the Merge Request.
3. CI/CD builds the containerized site and pushes the image to Docker Hub.
4. On the deployment machine, Watchtower detects the new image and rolls out the update automatically.

This removed the pain of frequent manual deployments and kept operations approachable — even a maintainer without deep system experience can ship changes, which makes handover and succession easy.

Key tech: Svelte (frontend) · Strapi (headless CMS — backend & database) · Git · CI/CD · Docker.

Operating under pressure

The site went live the same year I took over as lead, so every incident was uncharted territory — and, honestly, nerve-wracking. Through repeated firefighting I learned to stay calm: instead of panicking, I reason about the likely cause, then test and rule out possibilities in order until the problem is solved.

The scariest moments were the campus power outages — an unexploded bomb that could go off at any time. During the first one, the site appeared to run fine after the machines rebooted, but back-office data access was broken. I suspected the backend database container and tried to restart it — then Docker itself failed, and my heart raced (a data-security incident could have meant reporting to the Ministry of Education). The root cause turned out to be a disk read failure on the host machine, unrelated to anything I had done — but the lesson stuck. Since then, whenever the site needs repair I first switch it to a one-page static holding page I wrote, so visitors never see it visibly broken while the fix is in progress.