Jae Ung J.

Product manager for consumer mobile apps: subscription growth, retention, and shipping. Current client (US/UK wellness subscription app, hired here on Upwork, 8 months): App Store rating 2.0 to 4.5, trial to paid 10% to 17%, churn 8-10% to 2-3%. I also build and ship my own iOS apps with AI coding agents. Four are live on the App Store. Chalkak took 8 days from first commit to release, Fix My Korean 18 days in 13 languages. So I know what a spec costs to build, and when a developer says hard, I can tell hard from not-this-sprint. What I do for clients • Growth and retention for subscription apps: onboarding, paywall, trial conversion, churn, ratings, and the analytics to prove it (Mixpanel, BigQuery, RevenueCat). • Roadmap and specs written against the real codebase, with states and edge cases, so tickets are unambiguous. • Prototypes, not descriptions. I build the screens myself, then hand off to design and development. • Async first. Weekly written report with the numbers. Track record • Wellness subscription app (US/UK), Product Manager, January 2026 to now. Rating 2.0 to 4.5. Trial to paid 10% to 17%. Churn 8-10% to 2-3%. I run the AARRR analysis and coordinate the developer and designer. • LoveyDovey (AI character chat app), Product Owner. 10,000 to 2,000,000+ users in 11 months. Owned creator supply and discovery. Daily creator submissions 40 to 1,000+. • FODI (K-pop dance app), Product Manager. 2,000 to 20,000 users. Built the Mixpanel and BigQuery stack from zero. AI localization workflow cut ops effort by 95%+. • unglazed (my iOS studio): SAI Korean, Chalkak, Fix My Korean, Type Natural Korean. Evaluated the LLM grammar checker with a hand-authored 98-sentence golden set: 0 false positives, 96% accuracy. How I work Quantitative first to narrow the hypothesis, then interviews for the why. Reviews and support tickets feed both. Every recommendation comes with the metric that will judge it. Tools: Mixpanel, BigQuery, RevenueCat, Figma, Notion, Claude Code, Codex, Swift. Languages: English (fluent), Korean (native). Korean localization and QA on request. Share your app link and I'll bring a one-page teardown to the first call.