Designinginterfaces forAI agents

Product designer. Seven years in fintech and data-heavy products.

I’ve designed trading dashboards, payment flows and a data tokenization API — at Very Good Security in San Francisco, Hetman in New York, EstateLab in Norway. The problems I take have one shape: more data than fits on a screen, and someone who has to make a decision from it. I build design systems so the second feature costs less than the first. Now I design interfaces for AI agents, and use agents to build them.

Selected work

All cases

UnderstudyAI-native interfaces

An interview practice tool that shows whether an answer got better, not what it scored.

SIMCELSupply chain

A digital twin for supply chain planning — four scenarios compared in one view instead of rebuilt by hand in spreadsheets.

NFT MarketplaceCrypto

A marketplace for digital art built around two audiences that need opposite things — artists who need an audience, collectors who need to be first.

EstateLabReal estate

Cloud software that lets Norwegian real estate companies keep properties, contracts and tenants in one place — and see rental income, vacancy and expiry at every level.

How I work

Start with the constraint, not the canvas

I ask what breaks first: the data, the compliance rule, the legacy screen nobody can touch. Those limits shape the layout, so the design survives its first review.

Structure before surface

Flows, states and hierarchy get settled in low fidelity, where a change costs minutes. AI agents give me the first draft of that structure in a fraction of the usual time; refining it is still my job.

Handoff isn’t the finish line

I stay through the build — answering questions, fixing what drifts in the browser, checking the screen once more when real data finally lands in it.

The work I want next is AI-native — interfaces where an agent does part of the job and a person still has to accept or overrule the result. Write if you’d like to work together.

[email protected]

LinkedIn Behance