CS graduate who has done the IT-support-and-data day job and still makes small things for the web at night. By day: service management, reporting, documentation. After hours: the projects below.
The work here is modest on purpose; the point is to leave behind something useful, inspectable, or a little clearer than it was before.
Open to roles in operations, IT support, data, and administration — relocation abroad with visa sponsorship welcome.
The five projects below are the current public shelf. Each one has a different reason for existing; the descriptions stay short so the links can do the deeper talking.
| Project | What it is | Visit |
|---|---|---|
| Basirah — بَصِيرَة | A bilingual Islamic knowledge site for Myanmar Muslims, with selected Qur’an surahs, recitation audio, essential duas, and the Names of Allah in Burmese, English, and Arabic. | Live site · Source |
| Language Atlas | An offline-friendly field guide to 83 programming languages, organized for filtering and comparison across kind, paradigm, typing, execution mode, platform, and runtime. | Open the atlas · Source |
| LSTM Trend | A Streamlit forecasting experiment using Keras 3 on a PyTorch backend, with exploratory analysis, simpler baselines, interactive charts, and explicit limitations. | Try the app · Source |
| hobun-ssg | A learning project built around Hono and Bun: an app pre-rendered to static HTML and served on Cloudflare Pages. | Live site · Source |
| update-go | A small Bash installer and updater for the Go toolchain on Linux and macOS, with checksum verification, preview, self-update, and uninstall modes. | Read the script |
| Area | Tools that show up in the work |
|---|---|
| Web | Astro · SvelteKit · Hono · TypeScript · HTML · CSS |
| Runtime and delivery | Bun · Cloudflare Pages · Streamlit |
| Data and ML | Python · Keras · PyTorch · scikit-learn · yfinance · Power BI · SQL |
| Systems and tooling | Bash · Go · Git · Linux |
Build something small enough to inspect. Compare it with a simpler baseline. Keep the limitations visible.
That usually means a fast interface, a command that explains itself, a dataset with understandable categories, or documentation that says what a project does not prove.
These cards answer two different questions: how the profile is changing and which languages appear across its repositories. The activity card is live; the language card is kept locally so its palette stays intentional.
The repositories are the best place to see what I am exploring now. If something here is useful, feel free to borrow the idea, open an issue, or follow the thread back to the source.