UX Designer

Designing products that earn trust.

Decision-support tools, oversight interfaces, and the occasional working agent. The work is for moments where the wrong call has consequences, and where the system has to be explainable to the person standing behind the decision.

Selected Work
CAN-SIMPLAN — civic AI planning tool interface
AI Tool · UX Design · Design Sprint

CAN-SIMPLAN

In civic AI tools, accuracy isn't enough — the planner has to be able to explain the decision.
The Check-In Agent — behavioural fintech agent interface
AI Agent · Fintech · Behavioural Design

The Check-In Agent

First-time investors panic-sell at the bottom. This agent watches for the moment, and stays silent until it's earned the right to interrupt.
Clara — shared trip planning app with visible task ownership
UI/UX Design · iOS App Concept

Clara

An unassigned task is just the original problem with a due date. Clara makes ownership impossible to miss.
Conserva — premium food platform built system-first
UI Design · Brand System

Conserva

On a premium food platform, inconsistency reads as carelessness. Built system-first, tokens before components.
Aurafact — spatial archive of environmental sound
Product Design & Development

Aurafact

Environmental sound is a primary historical artifact — it just isn't kept. A spatial archive designed to change that.
Generative AI campaigns — stills and motion art directed by Yaya Falanga
AI Creative Direction · Stills & Motion

Generative AI

Self-initiated campaigns art directed from brief to final frame using Midjourney, Sora, and Grok Imagine.
About

I'm Jennifer Yaya, a UX designer based in Ottawa. I design AI products for high-stakes contexts—decision-support tools, oversight interfaces, and agents that have to know when to step back. My work sits at the layer where AI meets human accountability, which means most of it is about restraint: what the system shouldn't do, what it has to make legible, and where the person is still the one who has to defend the call.

The through-line is structural clarity. Making uncertainty visible instead of implied. Designing friction that prevents the wrong decision instead of dressing up the right one. Treating accountability as a load-bearing part of the interface, not a compliance footnote. The contexts vary (civic AI, fintech, productivity) but the discipline is the same.

I work fluently with AI as a build collaborator. WorkMap was developed end-to-end in React. The Check-In Agent runs on a system-prompted Claude endpoint I designed and wrote. I think out loud about the practice in Workingnotes. Outside of design, I host philosophy cafés, haunt bookshops, and think about sunny beach days more than is probably warranted.