Head of AI and builder, based in Saigon. Independently, I help builders and professionals ship real things with AI, and practise skills that actually stick.
AI productsBuilding with AIInterview prep, candidate-sideDaily practice systems
My path was not a straight line into tech: chemical engineering, then chip manufacturing at Intel, then automation testing at Katalon, then AI products. That mix is the point — I have shipped in factories, in test suites, and in AI products, so I care as much about what breaks in the real world as about what demos well.
I build AI products and lead AI work professionally. My consulting here is independent: what I offer is my own builder experience, in the open — including the public $0 to $1,000 MRR expedition documented on this same site.
The cross-industry builder: factory discipline, tester scepticism, AI speed. I did not start in AI. I switched into it — three times. Each switch looked like starting over. Each one actually added a layer I still use every day.
Engineering taught me to think in systems: inputs, processes, outputs, and safety margins. A small deviation upstream compounds downstream. The lesson I kept: do not optimise a step you have not understood, and measure what actually matters, not what is easiest to count.
In chip manufacturing, a tiny defect is not a rounding error. Yield, quality, and complex interdependent systems decide whether anything ships at all. The lesson I kept: quality is not a final check at the end. It is built into every step — and if you cannot measure it, you cannot trust it.
In automation testing, my job was to find failure modes early. "It runs" was never the same as "it works." The lesson I kept, and the one AI made urgent: code that runs can still fail — on security, on scale, on edge cases. Evaluating output is a skill in itself, not an afterthought.
Put those three layers together and you get how I build now: ship small, evaluate output against a clear standard, log the real mistakes, and turn recurring mistakes into rules. That is the Trace-to-Skill method behind PracticeLoop — a Skill Box built from your own errors, not from another course.
AI is changing workflows and roles fast. The defence is not knowing one tool. It is learning deliberately, every day, with reps — not just watching, but doing, checking, and fixing one thing at a time. That is what I teach, what I consult on, and what I am building in public on this site.
For: small teams and solo builders with an idea that needs to become a working v1.
What you get: a hard look at the idea, a free- or low-cost stack that fits it, and a plan for evaluating AI output — because code that runs can still fail, especially around security, scale, and edge cases.
For: builders who use AI tools daily and want better taste, not just faster output.
What you get: workflows, review habits, and prompt systems that catch AI's predictable blind spots, plus practical model routing on free tiers where it makes sense.
For: professionals in Southeast Asia preparing for English-language interviews for remote and global roles.
What you get: structured answers, English delivery, and daily practice reps — candidate-side only. I do not advise employers on hiring, ranking, or selection.
For: professionals learning English, coding with AI, or music — and anyone stuck watching courses instead of practising.
What you get: a floor-sized daily rep (10–15 minutes that counts), a log of your mistakes, and a Skill Box of rules distilled from those mistakes. This is the system behind PracticeLoop.
Pricing is on enquiry, scoped to the work — no packages invented here, no retainer pushed before we talk.
Start a conversation See the expedition →The fastest way is a direct message on any of these — I am Marcus N, @askmarcusn:
TikTok · Instagram · X · Facebook
Tell me, in one paragraph: what you're building or preparing for, where you're stuck, and what a good outcome looks like. I reply personally, usually within one to two business days.
Building in public: follow the $0 to $1,000 MRR expedition on the homepage of this site. All numbers there are true, starting from $0.