AI Voice companion for individuals with ADHD — designed and built the full product from scratch.

* Designed and implemented the entire front-end in TypeScript /  React, including 3D interactive scenes (Spline), dynamic UI components (shadcn, Tailwind), and responsive layouts — shipping code that matched the design vision with production-level polish.

* Architected the full-stack AI agent system single-handedly: React/TypeScript front-end, Supabase back-end (database, auth, real-time), and voice LLM-powered conversational AI pipeline — improving user focus recovery time by 75%.

* Architected end-to-end user flows from concept through interactive prototypes in Figma, then translated them directly into production code, eliminating design-to-dev handoff delays.

* Acquired paying users within the first week of launch, achieving an 85% weekly renewal rate — validating strong product-market fit and user retention.

* Confirmed commercial viability through seed user interviews (25% potential conversion rate) and organically grew a waitlist of 210+ users pre-launch.

* Directed the marketing lead on user testing and Wizard-of-Oz experiments to validate AI interactions before engineering investment. Authored SOPs for both engineering and marketing teams and defined success metrics to align cross-functional execution.

* Pitched at Berkeley SkyDeck and MiraclePlus (YC China); passed investor interviews.

Worldwide 1st Place (853 participants) — ElevenLabs × a16z Hackathon. Endorsed by CEOs of ElevenLabs and Vercel, and a16z partners.

* Led product design and front-end development of a multimodal AI travel agent that combines real-time voice interaction, dynamic route mapping, and location-aware storytelling — built and shipped as a fully working web app within 40 hours.

* Implemented the interactive UI in React with map integrations and voice-driven UX; collaborated with two engineers on LLM deployment via Dify to create a seamless conversational experience.

* Secured post-hackathon backing from Silicon Valley incubator Beta University, along with funding support from Microsoft for Startups Hub and Google — validating the product’s commercial potential beyond a prototype.

* Defined product strategy from 0-to-1: authored the PRD, scoped MVP features, and conducted on-site user research with 50+ real users to iterate on the experience before demo day.

Led design of Math Tutor product

Led the interaction design of an AI dating App targeting North American Gen Z:

* Independently designed the entire product interaction design, from 0 to 1. Through rapid user testing and iteration, and cost-benefit analysis of feature designs, successfully launched the iOS MVP within two months, achieving stable daily user growth.

* Conducted detailed user interviews and experience walkthroughs to identify user confusion with product features. Quickly updated improvement plans and product strategies to boost daily active users. Implemented agile development for rapid product iteration and A/B testing, achieving a 20% day-2 retention rate.

* Led the monetization efforts for the dating product. Through desk research, rapid prototyping, and prioritization, drove the launch of revenue-generating features.