AI Product & Growth2026 — Present
BeePOP: building the growth system for a commerce AI product
Connecting product development, SEO, measurement, commercialization, and overseas acquisition into one operating loop.
SELECTED CASES
Product case library
Practice across AI product development, overseas growth, SEO measurement, and global product systems.
View case
Overseas Tool Product2019 — Present
Weather: a mature tool product for global users
A focused weather utility shaped around real-time information, location, multi-city use, and reminders.
SEO Growth System2026
Anonymized SEO growth system: 929K impressions in three months
A measurement-led SEO system built around keyword strategy, content architecture, technical SEO, and iteration.
Global Product Infrastructure2022.03 — 2023.08
HelloTalk: real-time systems for tens of millions of users
High-concurrency foundations for global voice-room and real-time signaling experiences.
AI Product & Growth
BeePOP: building the growth system for a commerce AI product
AI Product Manager · Growth Lead · 2026 — Present
View caseThe product needed a repeatable way to move from AI capability and content demand to qualified users and commercial feedback.
Approach
- Connected UTM, GA4, and PostHog into one attribution loop.
- Built SEO/GEO, KOL, paid acquisition, and A/B testing workflows.
- Designed growth dashboards and operating rituals around activation and conversion.
Outcomes
- SEO/GEO external score reached 91.
- The product receives more than two GPT-referred users per week.
- ROAS reached 1.2 in the second month of paid acquisition.
Stack
GA4PostHogUTMSEO/GEOKOL GrowthFacebook Ads
Overseas Tool Product
Weather: a mature tool product for global users
Founder · Product & Growth · 2019 — Present
View caseThe product needed a clear utility loop and a reliable Google Play distribution path for global users.
Approach
- Designed the core utility flow around location and multi-city checks.
- Shipped the product to Google Play and iterated on overseas user experience.
Outcomes
- Built a complete tool-product loop from product design to overseas distribution.
Stack
AndroidGoogle PlayLocationWeather APIs
SEO Growth System
Anonymized SEO growth system: 929K impressions in three months
Growth Consultant · 2026
The goal was to turn organic search into a compounding acquisition channel without exposing the site identity.
Approach
- Mapped keyword intent to a repeatable content structure.
- Tracked search visibility, clicks, rankings, and conversion signals.
- Used the data loop to prioritize the next technical and editorial iteration.
Outcomes
- 929K impressions in three months.
- 5,724 clicks and an average position of 5.1.
Stack
Google Search ConsoleSEOGEOContent StrategyAnalytics
Global Product Infrastructure
HelloTalk: real-time systems for tens of millions of users
Golang Backend Engineer · 2022.03 — 2023.08
View caseGlobal users needed stable, low-latency communication across distributed services.
Approach
- Built Pulsar SDK and signaling services for real-time notifications.
- Designed Redis and MongoDB microservice patterns for global product traffic.
Outcomes
- Supported real-time signaling for tens of millions of users.
Stack
GolangPulsarRedisMongoDBMicroservices
Consumer AI Product
AI Health: a global consumer health product
Founder · Product Builder · 2024 — Present
View caseThe product needed to make an open-ended health experience feel understandable and useful.
Approach
- Designed a simple flow from user question to structured health context.
- Shipped and iterated through Google Play distribution.
Outcomes
- Built a working consumer AI product for global users.
Stack
AI AssistantAndroidGoogle PlayHealth Data
The product turns an open-ended conversation into a structured podcast session.
Approach
- Designed an agent-led interview flow for topic discovery and recording.
- Combined real-time voice infrastructure with guided product UX.
Outcomes
- Delivered an end-to-end real-time voice recording experience.
Stack
LiveKitRealtime VoiceAI AgentPodcast
Cross-border AI
Amazon Review Intelligence Agent
AI Product Manager · 2025
Sellers needed a faster way to detect recurring pain points across large review sets.
Approach
- Extracted context-specific feedback with LLM analysis.
- Generated visual reports for product and sourcing teams.
Outcomes
- Created a repeatable review-to-recommendation workflow.
Stack
LLMRAGDifyAmazon Reviews
AI Automation
AI Operations Agent and workflow automation
AI Product Manager · 2025
Operations teams needed repeatable workflows across content, research, and decision support.
Approach
- Mapped operational decisions into structured agent steps.
- Connected human review with automated execution.
Outcomes
- Prototyped a practical AI operations loop for commerce teams.
Stack
AI AgentWorkflow AutomationRAGPython