Minsun Kim
Sr. Product Manager - Enterprise AI Governance
Email:
Ambiguity, at enterprise scale, usually means everyone holds a different piece of the picture — the job is agreeing on what the whole one looks like. That means establishing a position, testing it against real evidence, and revising it publicly when the evidence changes. Conviction stays sized to what the evidence has actually shown.
Building AI systems draws on the same instinct, and it travels well: define what "safe enough" means for a given system, demonstrate it, and make it legible to the people who will ultimately own the outcome.
I learned this first by turning a role with no shape yet into one colleagues actively sought out — they moved from asking what the function did to asking to join it themselves. It's a shift the rest of the function, elsewhere, is still working toward. I measure it the same way now: not by what shipped, but by how much a team or a system can be trusted to run on its own. When that trust holds, it's not because I stepped back — it's because I built it to hold in the first place.
Minsun Kim
Sr. Product Manager - Enterprise AI Governance
Email:
Ambiguity, at enterprise scale, usually means everyone holds a different piece of the picture — the job is agreeing on what the whole one looks like. That means establishing a position, testing it against real evidence, and revising it publicly when the evidence changes. Conviction stays sized to what the evidence has actually shown.
Building AI systems draws on the same instinct, and it travels well: define what "safe enough" means for a given system, demonstrate it, and make it legible to the people who will ultimately own the outcome.
I learned this first by turning a role with no shape yet into one colleagues actively sought out — they moved from asking what the function did to asking to join it themselves. It's a shift the rest of the function, elsewhere, is still working toward. I measure it the same way now: not by what shipped, but by how much a team or a system can be trusted to run on its own. When that trust holds, it's not because I stepped back — it's because I built it to hold in the first place.
About Me
Healthcare, fintech, pharmaceuticals, and consumer platforms don't share much on paper — different regulators, different stakeholders, different countries. What connected them was the same underlying question, asked in a different accent each time: how do you get a complex system to a state where someone outside the room can verify that it holds up? A zero-to-one investment platform at LENDIT that lifted new customer acquisition 2.5% and traffic 5%. A clinician network of 3,200+ professionals at Johnson & Johnson. An SSO rollout at GSK that reached 80% of target healthcare professionals.
At Coca-Cola, that question took two different shapes. One was codifying the governance policy and Human-in-the-Loop review framework for an LLM-powered chatbot initiative — translating open questions about AI risk into structured documentation and review criteria a legal or security reviewer could actually use. That question stayed interesting enough to go build the rest independently afterward: risk scoring, policy generation, and runtime enforcement, coded and tested rather than assumed. It's on this site, working — not just described. The other was scaling a platform initiative across Korea and Japan, one country at a time, each with its own regulatory requirements and way of working.
What's next is less about the industry and more about the problem: complex systems where getting things right matters, and where the work has to hold up on its own, not because someone says it does.
Work Experience
April 2023 - Present
September 2020 - March 2022
November 2018 - April 2020
July 2016 - March 2018
The Coca-Cola Company - Technology Manager, Customer & Commercial
Codified governance policy and Human-in-the-Loop review criteria for an LLM-powered chatbot initiative, translating open AI risk questions into structured documentation. Owned product strategy and regional expansion of a Digital Asset Management platform, achieving 100% deployment across Korea and Japan within 12 months. Drove cross-functional technology delivery across Engineering, Data Cloud, and InfoSec, designing data integration and access governance with enterprise scale in mind. Built a Power BI–based catalog site from the ground up and automated approval workflows, report distribution, and data aggregation with Power Automate — replacing manual processes with self-service tools across the team.
GSK (GlaxoSmithKline) - Digital & Tech Specialist
Delivered GSK's first enterprise SSO initiative end-to-end, driving 80% adoption among healthcare professionals in partnership with engineering. Governed Technical Security Review (TSR) processes across internal and external digital services, aligning security, technology, and business requirements. Directed CRM and digital platform delivery with global engineering teams, translating business needs into scalable technology solutions.
LENDIT - Product Manager
Owned end-to-end strategy for a zero-to-one fintech investment platform, lifting customer acquisition 2.5% and site traffic 5% — a product that remained the company's core new-user acquisition engine for years after launch. Set the quality and release standard across the investment product lifecycle, from specification through production.
Johnson & Johnson - Digital Marketing Specialist
Built a zero-to-one clinician engagement platform reaching 3,200+ healthcare professionals across APAC, grounded in CRM insights and customer journey analytics. Instituted Digital Asset Risk Management and role-based access controls aligned to global compliance standards.