Shichen

STOCKHOLM · OPEN TO NEW ROLES

Hej, I'm Shichen.

AI Product Designer and Builder. I take products from the first question to a live app, and I design the AI part too: what the model decides, what the code decides, and the guardrails in between.

See the workDownload CV

Open to AI product design, product building and AI engineering roles in Stockholm.

idea to live app, solo
72 h
idea to live app, solo
list import, down from 10 min
1.5 s
list import, down from 10 min
fewer escalations at Christie's
40%
fewer escalations at Christie's
design tokens in one Figma system
454
design tokens in one Figma system

01 / SELECTED WORK

Four products, one thread: things people can trust.

LUS FOR LIFE · LIVE

LIVE

Group plans that actually happen.

A planner for friend groups where one person always ends up organising. Researched, specced, designed and shipped solo in 72 hours on Lovable.

  • Product design
  • Full-stack build
  • AI features
  • Brand
View case →

WANDERLUST · IN BETA

IN BETA

Your saved places, put to use.

An AI app that learns your taste from your Google Maps saves, finds you places at home every week, and plans trips anywhere. I rebuilt the backend over eight rounds, then pivoted the product on what users told me.

  • AI systems
  • Personalisation
  • Product pivot
View case →

CHRISTIE'S · POST-SALE

PILOTED

Making the wait visible.

Under 20% of clients finished checkout on their own. I found out why, designed and prototyped a status portal, and ran a pilot that became its business case.

  • Research
  • Experiments
  • Stakeholders
View case →
Image: HabitHero screens over token swatches

HABITHERO · DESIGN SYSTEM

CRAFT

Design craft, end to end.

Research, flows and a four-layer token system for a wellness rewards app. My most complete Figma build.

  • UI design
  • Design system
  • Usability testing
View case →

02 / HOW I BUILD

Five habits behind every product.

  1. 01

    Start from the number.

    A survey, a coded dataset, a timed test. I want to know the problem before I open a tool.

  2. 02

    Rules in code, judgement in the model.

    Opening hours, deadlines and travel time are code. Taste and wording go to the model. Faster, cheaper, fewer surprises.

  3. 03

    Guardrails where they hold.

    Access rules live in the database. AI output is schema-checked with a manual fallback. Suggestions come from lists I verified.

  4. 04

    Design the wait.

    In AI products, speed is a design problem. People need to see progress to trust it.

  5. 05

    Ship to real people early.

    Prototypes are polite. Real users are honest.

03 / NOTES

Thinking out loud.

I write about building with AI, and about making room for more women in it.

More on LinkedIn →
Photo: Shichen

04 / ABOUT

A little about me.

I started in marketing and auction operations at Christie's in London, running post-sale across 40+ markets. That's where I learned to frame problems with data and get Legal, Finance and engineering in the same room.

Then I moved to Stockholm and started building for what I was living: finding places that feel local, and finding my people. Those became WanderLust and Lus for Life.

I care about things people use for a long time and feel something about. Communities, gatherings, small rituals. I love meeting new people through meaningful craft and wellness events, and see new ideas get sparked up.

  • BSc Economics, UCL
  • MA Arts and Cultural Management, King's College London
  • UX and Product Design, King's College London

English and Chinese fluent · French intermediate · Swedish basic, getting there

Building something with AI in Stockholm? Let's talk.

Email meLinkedIn
GitHub →CV →