DeepSORT-Driven VIsual Tracking Approach for Gesture Recognition in Interactive System
Published by
IEEE
Work
Umi AI
|Product Designer
→
Summary
Led end-to-end product design for Umi AI, an app that enhances relationships through recorded conversation analysis
Highlights
Defined product flows, information architecture, and interaction models from scratch, translating abstract emotional insights into actionable and user-friendly interfaces
Collaborated closely with the founders and ML team to align the AI system’s capabilities with user needs, ensuring a warm, supportive, and non-intrusive product tone
→
Full-Time
Summary
Led the end-to-end design of Guidanz, a SaaS that helps technicians diagnose and repair heavy-duty engines, through building functions in web&mobile platforms.
Highlights
Designed online diagnostic session by reframing information architecture and introducing side panel feature, resulting in faster issue resolution times in field service operations
Built a scalable design system from 0 to 1, which reduces design-dev handoff time by 50% and ensured consistency across platforms.
Conducted on-site user research with certified Cummins technicians, gathering insights that directly informed design iterations and improved task completion rates.
Reconstructed the calibration installation process by streamlining the user flow and building an intuitive UI, improving task efficiency by 31.2%.
→
Intern
Summary
Led the exploration of design strategy and early-stage business use cases for Thunderbit.AI's MVP; partnered with the marketing team to grow DAU from 0 to 5K and supported the product monetization roadmap.
Highlights
Drove foundational research on AI productivity tools, including market and competitor analysis, to shape the MVP concept and core functionality within an AI-first product flow.
Led global expert and user interviews to generate key insights, define core user flows, and shape the product strategy from the ground up.
Introduced LLM-driven design thinking into UX strategy and implemented early-stage prompt design experiments to validate co-creation scenarios between users and AI agents.
→
Internship
Summary
Leveraged cutting-edge design practices to create experiences across both B2B and B2C contexts, such as e-commerce and data management.
Highlights
Shipped a live-streamed shopping mobile design (B2C) that drove 100% user satisfaction within 8 weeks.
Built main user flows for a global data management platform (B2B) by conducting research and incorporating UX strategy with business goals, resulting in a 20% rise in user adoption rates.
Facilitated cross-team meetings and coached clients and internal teams on Lean UX principles, ensuring timely design delivery.
Collaborated with UI experts to build the McKinsey Digital design system library, encompassing visual style, 50+ UI components, and 10+ interaction patterns.
→
Full-Time
Summary
Redesigned the onboarding experience of Lark, an all-in-one workspace tool. Led the user research for Lark for School MVP. Updated TikTok Trust & Safety guidelines through data-based user perceptions and expert research.
Highlights
Iterated Lark's onboarding and member invitation flows for both mobile and web, increasing new user engagement by 14% through user testing and pain points identification.
Led the on-site research for Lark for School by conducting quantitative research (surveys) and qualitative research (user interviews, competitive analysis), guiding the optimization of product designs.
Assessed and refined content moderation strategy by evaluating compliance measures and conducting 30+ global expert interviews and 10K+ user surveys across platforms.
→
Full-Time
Summary
Rebuilt MiFinance, a digital banking service, through iterating on the design of investment features.
Highlights
Optimized the account opening process of MiFinance by generating insights from customer feedback. Rebuilt the information architecture of the MiFinance dashboard, improving the user conversion rate by 25%.
Actively learned financial domain knowledge and regulatory compliance to better incorporate design into the financial experience.
Collaborated with the brand designer to build the Airstar Bank Global design system.
Education
University of Michigan
Master of Science in Information
Human Computer Interaction
Grade: 4.0
Pratt Institute
Bachelor of Fine Arts
Fashion Design
Publications
Skills
Design
UI/Interaction Design, Prototyping, Wireframe, Agile-Development, Design system, Motion Design, User Research, AB Testing, User Testing, Information Architecture, User Journey, User flow, Kano Model, Qualitative and Quantitative Research, Accessibility&Inclusivity.
AI Tool
Cursor, Bolt, Lovable, Notion AI, Uizard, UX Writng Assistant, Runway ML, Durable, Framer AI, Chatgpt, Deepseek, Midjourney, Adobe Firefly, Penpot+ AI, Thunderbit, Gamma.
Research
User Survey, Field Research, Heuristic evaluation, User interview, User testing, AB Testing, User Persona, Competitive Analysis(direct&indirect), Systematic thinking, Nielson Norman Group, User Journey Map, Eye tracking technoogy, Kano Model, Miro.
Development&Management
CSS, HTML, R, SQL, Python, Webflow, Agile Development, Cursor, Jira, Notion, Linear, Cross-team collaboration, Gronola.