Career Compass

These elements do not sit in a line. They form a gravitational system: experiences build capabilities, capabilities sharpen values, values shape direction. Click any node to see where it came from, what it means, and what it connects to.

The system

Connected nodes attract (spring force), all nodes repel (charge force). More connections = more central. Every edge is a verified causal link. Click to explore. Drag to rearrange.

Click any node to see its origin story, what it means, and every connection.
Connections
Values & identity
Capabilities
Experiences
Direction

How the layers formed

Each chapter added capabilities, shifted values, and sharpened the direction. The tags show what accumulated at each stage.

2014 - 2018
Seoul, University
Violinist turned textile engineer turned entrepreneur
Left music for something I could put into the world. Studied textile engineering and business. Started a sock company, failed, pivoted to baby socks, made 2.5M KRW from nothing. Learned that a garment work order and a product spec are the same skill: start from the person who will use it and work backwards. Discovered online platforms through selling, and that discovery led me to my first job.
0-to-1 instinct Failure as pivot, not endpoint Cross-domain thinking begins
2017 - 2020
Seoul, LG Fashion
E-commerce PM. 7 domains, 2-year call center embed.
Entered as an intern doing admin work. Taught myself SQL to work faster, then GA to build dashboards the team needed. Ran analytics workshops, became indispensable, went from intern to contract to special-hire (bypassing mandatory open recruitment). Managed 7 e-commerce domains end-to-end. Spent 2 years embedded in the customer service center. Watched users fail in real time. Turned that into 130+ product fixes, cut inquiry rate from 15% to 5%, generated $770K/mo. Shipped Korea's first ML virtual fitting service with a Berlin AI lab. Shown at CES. But commerce felt hollow. I wanted technology that does more than thin wallets.
Broke in: intern to special hire by creating my own role User observation instinct born Enterprise platform depth (7 domains) Self-taught: SQL, GA certification Early seed: tech should be genuinely useful
2020 - 2024
Seoul, Toss
Lead PM across growth, government, investment. Early member, 300 to 4,000.
Joined early because everyone should be able to manage money easily. Pushed retention from 70% to 85% through behavioral cohort segmentation. To truly fix finance, I had to fix the front door: customers needed government documents for financial services, and that process was broken. Led Toss to become a government-certified institution. Built the blockchain-based system with the government, launched Toss Residents' Center (official documents in-app). That unlocked digital ID, tax payments, fines, and a full digital government platform for 52M citizens. Presidential commendation. Pioneered investment advertising through regulatory sandbox. Youngest Domain Leader. Also learned what forced growth costs. Chose honest consent at MyData when every competitor used dark patterns. Eventually, the scope of what I wanted to build outgrew the company and the country. Left for the US to dream without domain limits.
52M-scale platform design Government certification + infrastructure User observation deepened (UT/IDI) Cross-domain: fashion to fintech to government Substance over foam crystallized (MyData) Growth over pride Discovered "users nobody designed for" Outgrew the domain
2024 - 2026
Boston / Global
Harvard MDE. 4 ventures. Scale AI. Kairos.
Zoomed out from Korea to the planet. Built AIRQUA (water from air, deployed in Thailand, Harvard President's Prize). Built ChoLab ($10 cholera kit replacing $1,000 lab tests, deployed in India, iF Gold, Dubai Future Foundation Top 100). Built Medly (safe OTC medication agent, MIT 1st Place). At Scale AI, analyzed 45+ frontier models. Discovered that benchmarks measure capability in isolation, not how models treat real people. That gap catalyzed everything that had been accumulating: user observation from LG, invisible-user awareness from Toss, the anti-manipulation conviction, the global lens from ventures. All of it crystallized into Kairos: someone who interprets technology must stand at the frontier and speak for those who cannot.
AI behavior evaluation Bicultural lens activated User observation applied globally Builder identity: beyond software Mission crystallized from all chapters Moon shot: write the standards

The thread through all of it

One pattern runs through every chapter: find the thing everyone assumed was stuck, break it open, make it feel obvious. The domains changed but the verb never did. The mission sharpened at each transition: "make useful technology" at LG became "fix the front door so finance works for everyone" at Toss, then "make AI treat people right" at Kairos. Each time, the ambition outgrew the container. Commerce felt hollow, so I moved to fintech. Fintech required fixing government, so I built the digital platform. Korea felt like a ceiling, so I left for the US. Capabilities stacked along the way: user observation gave ground truth, enterprise platforms gave scale credibility, cross-domain leaps gave range, AI evaluation gave the instrument. Values hardened through friction: substance over foam at Toss, growth over pride in confrontation, anti-manipulation chosen at MyData. Everything converges on one point: writing the standards for how AI treats people, and turning those standards into products.

Career Map

What I have, what I want, what I need. Where the experience flows and which paths lead to the destination.

Overview

Click any region. The border represents the environment I need. Outside is what I reject.

Click any region, the environment border, or the hate zone.

Where the experience flows

Click any node. Direct path skips Step 1.

Click any node to see details.
Toss
LG
AI
Harvard
Ventures
|
Bigtech
Startup
Fellow