About

Product leader, recovering test engineer, terminal-first operator. Based in Ho Chi Minh City. Writing in public so I stop repeating myself.

The short version

I'm Huy. I lead product for the AI initiative on the True Platform at Katalon — our next-generation platform that is repositioning the company from "automation" to "AI-native quality." Day to day, that means steering the Kai agent team (architects, engineers, QE) and shaping how our customers come to trust a machine that decides what to verify.

I spend my weekends writing, reading, and slowly rebuilding my daily workflow in the terminal with Claude Code. I believe the tools you choose reshape the thinking you're capable of. So I pay attention to them.

Where I started

I started as a manual test engineer. Automation back then was painful — clunky frameworks, overnight runs, brittle locators, whole teams held together by heroic regression sheets. I learned three things in that era that still shape my work today:

  1. The user's experience is the only real test case. You can pass every spec and still ship something nobody wants. Quality is a feeling before it's a metric.
  2. Scripts are a snapshot of a product that no longer exists. The second you stabilize a suite, the feature has shipped three versions. Chasing that treadmill taught me to aim higher — at intent, not steps.
  3. The best QEs are skeptics. Not cynics, skeptics — people who believe the product can work but want to see it prove it.

The pivot to product

Around 2017 I kept asking my PMs questions like "why are we building this" and "how do we know it's good" — eventually someone said, "you should just do that job." So I did. I moved into product management and spent the next stretch of my career learning that "shipping well" is mostly about rituals, calm, and honest conversations — not brilliant ideas.

I've led product at several stages: early-stage discovery, platform bets, integration initiatives, and company-wide repositioning. The thread through all of it: I care a lot about the loop — from customer signal to decision to shipped code to real outcome — and I get impatient when a team's loop is longer than it needs to be.

What I'm doing now

I'm leading Katalon's AI initiative — a set of AI-native capabilities inside the True Platform, including the Kai agent, AI Test Generator, AI Test Runner, AI Insights, AI Briefing, and Bug Reporter. The bet: software quality is about to change shape entirely, and the companies that rethink the whole workflow (not just bolt AI onto old flows) will lead.

My core belief, in one sentence: Agentic QA is where the agent holds the product's intent — not a test script — and decides what to verify against it. Everything else is commentary.

I also run a personal agent system — an Obsidian vault plus Claude Code plus a handful of custom skills — that gives me a daily briefing, processes meeting transcripts, writes first drafts, and keeps a living CRM of the people I collaborate with. It's not a gimmick; it's genuinely how I work now. I wrote up how to build one here.

How I try to work

  • Listen until the problem is clear. Rushing to build is how teams end up with beautiful answers to wrong questions.
  • Ship in tight loops. A two-week spike you can throw away beats a six-month roadmap you have to defend.
  • Keep the pressure manageable. Panic is not a strategy. Calm teams ship more.
  • Tell the story honestly. Leaders want fewer surprises. Peers want less theater. Customers want the truth. Same answer fits all three.
  • Write it down. If a decision isn't written, it didn't happen. If a principle isn't written, it's just a mood.

What I care about outside work

  • Reading late into the night. Mostly operating manuals for humans — biographies, essays on craft, the occasional weird novel.
  • Travel with my wife and kids. The pace of life is a feature; I want to build one that feels generous, not rushed.
  • The quiet side of tech. Terminals, keyboards, plain text, scripts that actually work. The craftsperson's side of the industry, not the hype side.

How to get a hold of me

If you read something here that rang true, or if you're working on something weird in the agentic QA space, I'd love to compare notes.

Or, if you have a specific thing you want to bounce around, pick my brain →