Explainability before accuracy.
Enterprises adopt AI security tools in proportion to how much they trust them — not how accurate they are. Ship the explanation before the accuracy improvement.
The Chain
Four layers, run in order. Skip one and the next still exists — but without the grounding that makes it reliable.
- 01
Philosophy
Why the product exists — the highest-order decision criterion.
- 02
Principles
Repeatable decision logic derived from philosophy.
- 03
Conceptual Models
Durable abstractions that make individual feature decisions predictable.
- 04
Execution
Stage-based workflow that operationalizes the layers above.
Each layer is load-bearing.
Operating Philosophy
Enterprise AI security is a negotiation among trust, speed, and scale. Four pillars keep the negotiation honest.
Day 2 is the real launch.
The critical product phase begins after the contract is signed. Drift detection and silent operator reliability decide the renewal — not the demo.
Adversarial thinking by default.
Before shipping a feature, ask: how would I exploit my own AI to leak data? Guardrails are what let the enterprise accelerate without risk exposure.
Translate risk into value.
Speak the language of the CISO, the CFO, and the developer at once. Roadmaps that speak in outcomes earn executive trust faster than roadmaps that speak in features.
The Discipline of No
The hardest part of enterprise product management is not deciding what to build — it is the discipline to refuse what does not belong.
The four costs of scope creep- Identity dilution
The product stops being the best at one thing and becomes average at many.
- Wrong competitive set
New features drag you into markets where you have no advantage.
- Execution fragmentation
Team focus splits; quality degrades across both domains.
- Workflow mismatch
Features built for one persona create confusion for another.
Saying no does not mean the need is invalid. It means this product is not the right solution for it — and pointing to the one that is.
Selected Work
Three chapters of enterprise product leadership — zero-to-one, portfolio scale, and personal craft.
- 01
Samsung Knox Enterprise Security — AI & Zero Trust
Director, Strategic Product Management · Samsung Research America, 2024–presentDirecting Samsung Knox’s enterprise security portfolio — Zero Trust architecture, on-device AI governance, and FedRAMP readiness — across engineering, business, and regulatory domains.
150M+Galaxy devices under Samsung Knox platform coverage15%Knox solution MAU increase - 02
Samsung Knox Suite
Global Lead Product Manager · Samsung MX, 2018–2024Took Samsung’s enterprise mobility SaaS bundle from 0 to 1 across 30+ markets — EMM/UEM on a platform that secures over two billion devices worldwide.
$200M+enterprise SaaS & services ARR11.5Mannual enterprise device sales55Kbusiness customers30+markets launched - 03
PM Intelligence System
Independent project · 2024–presentA production-grade AI operating system for PM workflow — multi-agent orchestration, RAG-based knowledge retrieval, and human-in-the-loop decision gates. The proof is the site itself.
1person, end-to-end — product, architecture, design, engineering, and operations
Trajectory
Architect first, product leader second — in that order, on purpose.
- 2008–2011POSCO DXSoftware Architect, Platform R&D Center
- 2011–2014Samsung Mobile eXperience (MX)Software Architect & Technical Partnership, Samsung Smart School
- 2014–2018Samsung Mobile eXperience (MX)Product Strategy & Technical GTM Manager, Device Software Customization
- 2018–2024Samsung Mobile eXperience (MX)Global Lead Product Manager, Enterprise Mobility Platform & SaaS
- 2024–presentSamsung Research AmericaDirector, Strategic Product Management, Enterprise Mobile & AI Security
- Carnegie Mellon University · MSIT, Software Engineering
- Korea Advanced Institute of Science and Technology (KAIST) · MS, Software Technology