中文 / EN
Experience
Back to home

Professional experience.

Product leadership across digital health, insurance, and healthcare platforms.

Northeastern University

Teaching Assistant for CS5001 · Vancouver, BC · Jan 2024 - Apr 2024

  • Delivered technical guidance on algorithm design, complexity analysis, and debugging in Python-based computational assignments.

CinoCore

Spun off from LinkDoc

Product Manager · Shanghai, China · Oct 2021 - Oct 2022

  • Owned the product roadmap for a post-surgery pulmonary rehabilitation Digital Therapeutics (DTx) platform, productizing clinical follow-up workflows and medication adherence management.
  • Led cross-functional collaboration with engineering, clinicians, and hospital operations; translated clinical needs into product specs, user stories, and sprint-ready requirements.
  • Contributed to securing China’s first NMPA-approved oncology DTx; drove clinical deployment across multiple top-tier hospitals, serving 5,000+ patients annually.

TH-002 6.0 Major Release

Mar 2022 - Jun 2022

Goal Drive a strategic shift from “reactive management” to a “compliance-driven model,” improving patient engagement and product satisfaction while strengthening automated intervention.
Responsibilities Built and validated six core hypotheses around patient learning curves, feedback perception, and family compliance; redesigned patient-side task and incentive logic, strengthened real-time feedback, and optimized clinician-side automated intervention paths.
Outcomes
  • Significant increase in post-discharge compliance; manual follow-up satisfaction reached 80%+.
  • Product capability shifted toward automation; frequency of manual intervention by clinicians and ops dropped, with first manual intervention delayed by an average of 3 days.

ZhongAn

Online P&C insurance company

Senior Product Specialist · Shanghai, China · Nov 2020 - Aug 2021

  • Owned the direct-channel accident and property insurance product lines; transformed complex insurance bundles into an intuitive user decision framework, launched 8 new products in 6 months, and drove 86% YoY premium growth through portfolio upgrades, new sales model design, vertical category channel launches, and external traffic acquisition optimization.
  • Redesigned core user journeys (registration/login, conversion funnel, claims experience) on the ZhongAn Insurance App to improve conversion efficiency.

App Onboarding & First-login Conversion Optimization

Mar 2021 - Aug 2021

Goal Improve activation and login conversion while strengthening user data capture for downstream conversion and lifecycle marketing.
Responsibilities Identified low-login segments from app-store acquisition; redesigned onboarding to include a pre-login risk assessment module, reduced friction in login UX, and introduced onboarding offers to incentivize registration and first login.
Outcomes The risk-assessment flow contributed ~4.5% of daily new first-logins; telesales follow-ups drove ~35k RMB/week incremental premium.

Tencent Trusted Doctors

Acquired by Future Doctor

Product Specialist · Shanghai, China · Jun 2019 - Nov 2020

  • Led product planning and iteration for C-end healthcare service platforms (WeChat Mini Program + Web), driving registered users past 1M through user growth strategies and product optimization.
  • Built CRM, marketing, and membership systems from 0 to 1, enabling automated onboarding for 30+ business partners and reducing integration effort by two-thirds.

ChatSprout 2.0 — LLM-Powered Scenario-Based English Communication Coach

Nov 2025 - Mar 2026

Goal Identify the core pain point: non-native newcomers struggle not with grammar, but with context-specific expression in high-stakes scenarios (networking, disagreeing, giving feedback).
Approach Conducted 10+ user interviews and designed a scenario-based training system using LLM + RAG — built a scenario taxonomy from user research, with a low-similarity fallback mechanism to handle edge cases and maintain user trust.
Results Optimized the retrieval pipeline to achieve ~80% Top-3 scenario match accuracy; established an evaluation framework (expression accuracy, tone, latency, token cost) and controlled the response time in 1.5–2s.
Key insight In open-ended language training, retrieval quality drives generation quality more than model scale.

LAIX Inc. (Liulishuo)

Product Specialist · Shanghai, China · Apr 2017 - Jun 2019

  • Scenario-Based Spoken English Course: defined the product format and pre-sales funnel through competitive research; scaled WeChat ecosystem acquisition via distribution and paid channels, improving pre-sales conversion by 5x through A/B testing.
  • IELTS Liulishuo: led 0 to 1 commercialization, owned traffic acquisition, conversion operations, data analysis, and community SOP design across the App and WeChat ecosystem.