01 — The problem
Most AI-generated portfolios look fast.
They also look generic.
The goal wasn't speed.
It was to build something deliberate — under time pressure.
02 — The approach
AI wasn't the designer.
It was the accelerator.
I used Claude to generate, iterate, and build faster —
but kept control over structure, decisions, and quality.
- generate → evaluate → refine → repeat
- accept what works
- reject what feels generic
03 — What I built
A portfolio structured as:
- an immersive entry (spatial hero)
- narrative-driven case studies
- a consistent but flexible system
04 — Outcome
Designed and shipped in 10 days.
Fast — but controlled.
Efficient — without losing authorship.
05 — Takeaway
AI speeds up execution.
It doesn't replace judgment.