Hi! I’m Yusen,

An AI product designer who drives products forward with technology.

  1. 2026

    Carnegie MellonMHCI

  2. 2022

    MetaiOS & Web Engineer

  3. 2020

    MicrosoftWeb Engineer

  4. 2019

    AmazoniOS & Web Engineer

Portfolio

AI-powered workflow · Interaction design · User research

Designed and delivered a voice-driven agent for warehouse processing

An 8-month CMU MHCI capstone with Goodwill of Southwestern Pennsylvania, whose 32-store softlines business brings in about $23.66M in annual sales. On the team, I led product design and was the tech lead. Starting from user research, we built Goodwill’s first internal AI workflow tool. Processors say a garment’s attributes, and the agent operates the existing POS system for them to print the price tag. In a one-week in-store pilot, a processor went from about 39 to about 83 soft-line items an hour, a 112% increase.

Goodwill Agent voice strip docked over the existing POS system

Social networking app · Multi-device experience · Interaction design

In-person meetups for six, designed for Bumble BFF

An individual project in CMU’s Advanced Interaction Design course. The brief asked me to raise conversion and engagement on Bumble BFF, a major US app for making friends. I designed a new feature that hosts in-person meetups for six: each participant has dinner or another activity with five strangers, at a venue the system books in advance. The experience covers iOS, Android, Apple Watch and a web entry point.

Three Hive Five screens: this week’s events, an upcoming dinner counting down, and the booking confirmation

Vehicle HMI · Physical controls and dashboard design · User research and validation

Semi-Autonomous Vehicle Handovers a rider can trust

A team project in a CMU interaction design course. We designed the physical controls and digital dashboard of a semi-autonomous three-wheel electric moped for a rider with limited mobility. Based on usability testing with 6 participants, we refined the three interactions that matter most: switching between manual and auto mode, getting help in an emergency, and handing control back to the rider when the vehicle can no longer drive itself safely.

The auto-mode dashboard: a 3D view from behind the moped, a speed dial at 30 mph, and the prompt Vehicle ahead, maintaining distance

At Meta · iOS feature design and delivery

Clickable Category on Instagram

As the lead iOS engineer, I drove the launch of the Clickable Category feature on Instagram creator profiles. In an experiment covering 40M+ users, the feature achieved 8% CTR, and creators saw +9.3% MoM follower growth.

The Clickable Category feature on an Instagram profile