Boat Refugee → Harvard Economics → 30+ Enterprise Software → FDA Expert → CEO/Founder
Born in Vietnam in 1976. Age four, family escaped by boat. Arrived in America as refugees with nothing except skills and work ethic. Dangerous crossing with no guarantee of safe harbor.
Foundation: A life characterized by building toward better conditions rather than accepting constraints of present ones.
Self-taught coder since childhood. Won Boston citywide coding contest in 4th grade. CS50 with Brian Kernighan — co-author of “The C Programming Language,” creator of “hello, world.”
Kernighan’s principles: Code should be clear, not clever. Programs should be small and do one thing well. Documentation readable by humans.
Economics framework + self-taught coding = the intellectual DNA of Stillwater.
Built the eSource platform: electronic capture of clinical trial data at point of care, with complete audit trail from moment of capture. No transcription. No paper intermediary. The original electronic record is the source document.
Outcomes at scale:
The Foundational Insight:
“In clinical trials, ‘trust me’ is not evidence. Only the original, timestamped, attributable record is evidence.”
This sentence is the intellectual seed of Stillwater. Transplanted from clinical data to AI outputs: LLM prose is not evidence. Only the original artifact, captured at execution time, with hash-linked audit trail and attribution to authorizing principal, is evidence.
The translation from clinical trials to AI verification is not metaphorical. It is architectural.
FDA 21 CFR Part 11 (electronic records) → OAuth3 + rung system
ALCOA data integrity standard → Evidence bundles (Lane A / B / C)
Clinical trial phases → Verification rungs (641 / 274177 / 65537)
Every design choice in Stillwater has a CRIO ancestor.
Verification architecture is not a constraint on capable systems. It is the foundation of trustworthy ones.
Open-sourced Stillwater, the verification OS for AI agents. Built Solace AGI on Part 11 architecture.
Question: What if every AI action was logged, auditable, and verifiable like a clinical trial?
Answer: Solace. ALCOA-compliant evidence. OAuth3 agency tokens. Hash-chained audit trails. Verifiable outcomes.
Not just safer AI. Enterprise-grade, auditable, trustworthy AI agents.
Fail-closed, never fail-open. Block rather than guess.
Every action logged, auditable, explainable. No black boxes.
Built for the hardest problem. Everything else is easier.
Proven in 360+ clinical trials. Not theoretical.
I’ve spent 10+ years building enterprise software and safety-critical systems. This is my bet that we can build AI agents that are safe, auditable, and actually useful.
When you use Solace, you’re not just getting a tool. You’re getting hard-earned lessons from clinical trials, audits, and regulatory scrutiny.
