🚧 Draft / coming soon. This is a placeholder for a full write-up. The notes below are the outline I want to flesh out — not the finished piece yet.
This is the first entry in Experiments — where I share what I'm learning, tools I'm trying, and things I'm building in the open. Fittingly, the first experiment is this very site.
I built this portfolio with Claude Code as the coding agent and FireCrawl to import the articles I've written for clients. Here's what I want to cover when I write this up:
What I want to write about
- The idea — why I wanted a portfolio that doubles as a searchable archive of my client work, not just a list of links.
- The import pipeline — using FireCrawl to scrape articles into MDX, then the cleanup work to strip each publisher's page chrome (and the surprises along the way, like dates hiding in schema.org instead of meta tags).
- Working with an AI agent — where it sped things up, where I had to steer it, and the moments it caught bugs I would've missed.
- Design decisions — the editorial dark theme, centralizing colors as tokens, and rendering real client logos.
- What I'd do differently — the honest version.
More soon. If you want the short version before I finish writing, reach out.