none of this
was necessary._

I wanted to actually see where my money goes, so I built this. It pulls every transaction from my Up Bank account, categorises them, and renders the whole picture through custom D3 charts.

Go backend, PostgreSQL, vanilla JS frontend. There's a Sankey flow chart that shows spending by category, a streamgraph for net worth over time where you can click into individual bands, and a review queue for pending transactions. The categorisation uses pattern matching and an LLM for receipt extraction. Both were overkill. I'd do it again.

3,800+ transactions categorised across 3 years of real spending data. Not a side project I abandoned — I use it every day.

Everything I build ends up running on a single Intel NUC sitting in my room. Caddy handles routing, Docker Compose runs each service, and GitHub Actions deploys everything automatically when I push.

I built a site onboarding system so adding a new project is just a YAML file and a PR. Deploys build the image, push to GHCR, trigger the NUC to pull and swap the container, and roll back if something breaks.

Cloudflare acts as a TLS-terminating edge — same role as an AWS ALB — with the tunnel handling the encrypted leg back to the NUC. Caddy routes by hostname behind it. This site is running on it right now.

My mate Norton wanted to build a florist discovery site for his girlfriend. I said I'd help. Ended up doing the architecture, security hardening, deploy pipeline, and hosting it on my NUC.

Next.js, TypeScript, Supabase, Tailwind. Search page with filters, florist profiles, enquiry form. I went through and hardened the API routes, added error handling, and made sure Supabase wasn't leaking internals to the client.

His project, his repo. I just made sure it wouldn't fall over. It will fall over. Norton must really love his girlfriend.

First actual client project. DeltaLilac is an artist who needed a proper home for their work, something more than a Twitter feed or a Linktree.

Full artwork database with a featured showcase on the front page, commission status toggle, and light/dark mode. Next.js, PostgreSQL, Tailwind. Self-hosted on the NUC like everything else.

Still in progress. The database schema is solid, the showcase works, the dark mode is suspiciously good for a first pass.

Automated job application pipeline across four sources — LinkedIn, Seek, Indeed, and Glassdoor. Each listing lands in a SQLite store, gets scored by Claude against a target profile, then passes through a tailoring pipeline that generates a bespoke resume and cover letter for each posting.

Python throughout. WeasyPrint renders the final PDFs. Playwright handles browser automation for submission. A FastMCP server exposes the pipeline to Claude Desktop for manual intervention. Runs on a daily NUC cron with a dry-run gate to prevent accidental mass-applies. Application status tracked in a local dashboard.

Private repo — it knows too much about my job hunt to be public.