Product Lead (PMF experiment)

• Shipped canon-faithful AI companions on WeChat and iMessage in 2 days; week 1: 430+ organic users, 200K messages, and 40%+ D7 retention at $0 spend (top character: 564 msgs/user).
• Wrote the base persona prompts and designed the character voices and voice-trigger logic (when a companion replies in voice instead of text), tuned for emotional delivery.
• Designed the behaviors that make companions feel alive: proactive messaging, time awareness, short-term and cross-platform memory; specs grounded in the engine's codebase.
• Ran GTM, community, and support end to end; the launch post went viral and the community hit 100+ members in week 1, hundreds of messages a day, with bugs and feature requests coming to us directly.

Founding Product Designer

• Built a cross-product, cross-device design system from scratch, speeding cross-functional alignment and iteration.
• Shipped a 0 → 1 AI customer-support system (RAG + LLM) with grounded citations, an answer/route/escalate taxonomy, and a human-in-the-loop review queue, now live with daily results logged to a Notion database to track deflection.
• Owned the 0 → 1 design of a merchant operations app with two AI workflows and guardrails (status-aware eligibility, audit trail); shipped iOS/Android with 3x growth in 2 months and 20% fewer support tickets.
• Rebuilt a complex B2B onboarding flow with AI-accelerated prototyping, cutting setup effort ~30% and aligning PM and engineering in days.
• Overhauled the UX of a high-volume payment app (8,000+ terminals, 400K+ weekly transactions), cutting operational errors with role-based UI and state-driven interactions.
• Worked at founding-team scope: partnered with the CEO on strategy and investor decks, and shaped GTM and enablement materials.

Product Design Intern

• Researched and designed vehicle lighting interactions for an L2/L4 autonomous truck convoy with no industry precedent; AI-assisted prototyping cut testing iteration 50%.
• Designed an in-vehicle HMI for lane-change warnings, improving signal noticeability 15% with dynamic risk cues that don't obscure core driving info.
• Built an internal dispatch platform replacing manual coordination, cutting operational turnaround time 80%.

Product Design Intern

• Led research and usability testing for a consumer mobile AI gift-shopping assistant: surveys, competitive analysis, head-to-head layout tests, and A/B testing across 100+ hi-fi screens.
• Iterated from test results to +35% purchase-step conversion and +20% task completion.
• Tuned the assistant's conversational behavior, including when it answers directly versus asks a clarifying question, to keep responses predictable across open-ended inputs.