I build AI agents, CRMs, analytics pipelines, and compliance gates for medical, beauty, patient-acquisition, and data-heavy operations. The work is shown with screenshots, implementation detail, and measurable system evidence.
Case studies organized by skill, implementation, and screen evidence
9/9
A patient-acquisition platform that connects Korean hospitals, treatment packages, language support, pricing, consultation intake, and partner/provider case operations.
An AI marketing operations console that coordinates ad performance, consultation funnels, content calendars, competitor pricing, and medical-ad-law gates across four clinic branches.
A 4-branch analytics platform reconciling 17 months and 23,555 revenue rows into channel attribution, booking funnels, and no-show risk views.
A 256-page mobile-first fortune-telling platform with Toss payments, coin economy, idempotent refunds, classical-text RAG, and push dispatch.
A TypeScript monorepo that runs crawl, analyze, workorder, and recheck loops deterministically without LLM dependence, isolating AI inside a replaceable ai-core.
An operations knowledge base that structures medical-ad-law rules, internal guidance, and raw sources behind PII defense and source immutability.
A security ingestion gate that re-scans PII regardless of source declarations, quarantines uncertainty, and never logs sensitive values.
A premium mobile saju SaaS that grounds OpenAI readings with classical citations and experiments with membership, coins, and collectible reports.
A suite of PowerShell, Python, and Apps Script tools for file sorting, PC health checks, daily CRM reports, and blog monitoring.
I build products and automation for industries where mistakes are expensive: medical, beauty, foreign patient acquisition, and operations-heavy teams. I prioritize working consoles, CRMs, price monitoring, content gates, and patient journey systems over decorative landing pages.
My operating rule is simple: legal and privacy risk should not depend on human attention. Medical-advertising copy is blocked before publish, patient PII is isolated behind layered defenses, and revenue/funnel reporting is kept reproducible through idempotent pipelines.
This portfolio collects the strongest systems I have built so far, with screenshots, explicit skills, implementation notes, and the numbers that can be verified from the work.