Portfolio 2025

A single dark page against a night sky, split between a short bio on the left and featured projects on the right, with the work filed under four disciplines: website development, web and mobile app development, UI and UX, and passion projects. Next 15 and React 19, with Framer Motion, tsparticles and Radix behind it.

It was built to feel like something. A CRT scanline overlay, a fluid cursor trailing the pointer, a custom cursor ring, drifting particles, a welcome modal, a loading overlay and an animated overlay between theme changes. Every one of those was a deliberate choice at the time.

Rebuilding in 2026 I cut every effect listed above, and the principle that replaced them is the first line of the new site's design spec: restraint over decoration. Two interactions survive there, and both are evidence for a claim a case study makes rather than atmosphere for its own sake.

The honest reason is that the effects were doing work the writing should have been doing. A hiring manager reads a portfolio for judgment, and a fluid cursor is not judgment. Once the case studies actually argued something, the scanlines had nothing left to carry and their cost, in bundle size and in attention, stopped being worth paying.

I keep this one online because the comparison is the point. It is the same person, the same work, and a different answer about what a portfolio is for.

The dependency list outran the site. The OpenAI and Vercel AI SDKs are installed and imported nowhere, left over from a feature I planned and never built. Shipping a package you do not use costs install time and audit noise, and it makes the manifest a bad description of the project.

Filing work under four discipline categories was the other mistake. It let me show everything, which meant a visitor had to do the sorting. The rebuild shows five case studies and puts the rest in an archive, which is this list.