What I work on
I am Cao Hung Nguyen. The AI Whisperer is where I examine how conversational AI can develop a recognizable character without pretending to be human. My focus is AI personality engineering: identity, voice, memory, continuity, and boundaries across conversations.
I came to this work through building AI systems that people actually use. Rintori connects recordings and personal notes to a conversational assistant that can point back to its source. SkillDiscs turns source material into learning paths with an AI tutor. Boniforce makes business-risk analysis more transparent. These are different products, but each has taught me something about the same problem: an answer is only useful when the system can explain where it came from and respect the person relying on it.
Why personality needs engineering
A short prompt can make an assistant sound warm. It cannot, by itself, decide what the assistant should remember, how it should respond when challenged, or whether the same character survives a long conversation. I want to make those decisions explicit and testable. That includes the harder questions: false memories, excessive agreement, emotional overreach, and the difference between a useful companion and a manipulative one.
My working principle is simple: believability means coherent behavior, not deception. An AI can be responsive and recognizable while being clear that it is an AI.
What you will find here
I write about conversational AI, AI companions, memory and personality continuity, evaluation methods, and responsible human–AI interaction. The AI personality engineering guide explains the framework behind this work. The articles page collects current writing and keeps the older German archive accessible. The German transition note explains why this publication is moving into English.
My book AIXec covers the broader discipline of implementing AI systems in practice. It is part of my technical background, not a claim that every production challenge is a personality challenge.
Start a conversation
If you are thinking through an AI personality, a companion, or a conversational product, send me a message. Tell me what you are building and where you are stuck. I would rather start with a specific problem than a generic pitch.