Hi, I'm Mason —
I write, I translate, I ship real AI solutions

Over the past 8 years, I wrote more than 1,600 articles — 1,000+ published under my byline, 30+ each exceeding a million pageviews. My career started in 2018 as a Japanese translator/editor at 早安健康, adapting Japanese health media content into Traditional Chinese. A year later I went fully freelance, continuing to contribute to 早安健康 on contract while steadily building a roster of long-term clients across the region — MedNet (an online doctor consultation platform) in Taiwan, ANKH Functional Regeneration in Hong Kong, and Hakuhodo DAC Taiwan and MyBest Inc. from Japan. Every day of those 8 years involved reading English and Japanese medical journals, cross-language adaptation, and long-form SEO writing.

The most important thing those years taught me: clients never pay for "writing skill" — they pay for "you understood the problem I couldn't articulate, and then you articulated it for me." That's exactly why I believe I'm a natural fit for FDE/SE roles — I've been doing precisely this for 8 years in a "non-engineering" domain. The only difference now is that I also have the technical ability to demo and ship on the spot, right in front of clients.

In 2024, when I first seriously used ChatGPT, I had a realization: my bottleneck for the past 8 years was never a lack of ideas — it was a lack of an execution layer. I'd written countless articles, but actually "building a working website" required crossing a gap I'd never crossed — code, toolchains, deployment, infrastructure. That gap swallowed too many of my ideas. AI filled that gap — not because it taught me to code (frankly, I still don't really understand the code), but because it became my execution layer. I'm responsible for knowing exactly what I want, describing it precisely enough, judging whether the output is correct, and iterating until it actually works. AI handles turning my thinking into running code. After two years of this collaboration and verification process, in 2026 I shipped the site you're looking at — 195 AI articles, a complete tech stack, an automation toolchain, all planned and verified by one person.

This experience changed my definition of "capability." We used to believe capability came from degrees, seniority, and training — to build a decent website, you'd first need three years learning to code. But in the AI era, capability is a different set of dimensions: Can you articulate what you want? Can you describe it precisely enough? Can you judge whether the output is correct? Can you iterate until it actually works? These four things are harder and more valuable than "learning the syntax of N programming languages," and AI currently can't do them — it needs a person with judgment to guide it. That's exactly the foundation my 8 years of content work gave me: I've always been doing "understanding what someone can't articulate, then articulating it for them." The only difference is that before, I did it for clients; now I do it for myself; and perhaps next, I'll do it for an AI company's clients.

This energy — I might bring it to a real AI company, combining 8 years of "understanding what clients actually mean" and "writing content with real reach" with a full year of hands-on AI collaboration to build an entire knowledge platform, becoming the person who can understand requirements on the spot in a client meeting, demo with AI in real time, and verify on the spot whether it can actually ship. Or I might continue as an independent consultant, bringing this "product judgment x AI execution" combination to more clients who need it. I'm walking both paths. Whichever conversation happens first.

There's also a bigger context here — this isn't just my personal bet. NVIDIA CEO Jensen Huang publicly said "AI is the revenge of the English Major" — he believes the most valuable skills in the AI era are language, writing, creativity, storytelling, and understanding the human condition, not more programmers. That quote lands even harder when you consider my 8 years of content work. And for an audience reading this in English: this is your native advantage. But here's the twist — I'm proof it works across languages too.

The market is validating this in real time: Anthropic, OpenAI, Palantir, Salesforce, and Cohere are all expanding their Forward Deployed Engineer teams. In 2025, demand for these roles grew 800% year-over-year — and the core competency is "can listen to clients, can translate ambiguous requirements into product, can collaborate on the spot." Soft skills are outweighing pure technical skills. Meanwhile, Replit, Cursor, bolt.new, Lovable — an entire generation of tools exists to test the proposition "can someone with zero coding background actually ship real products."

I should add something important: the way I think is very different from a traditional engineer's, and we're each good at completely different things. Writing high-performance distributed systems, designing elegant data structures, debugging race conditions at 4 AM, whiteboard algorithm complexity analysis — none of that is my domain, and I'm not going to pretend otherwise. But what an AI company sometimes needs isn't yet another person who thinks the same way as the existing engineering team — it's a different angle of looking at things. Cognitive diversity is diversity too — and this might be exactly the piece I can contribute.

Following that thread of diversity, there's a practical question you're almost certainly asking: my English reading and writing are solid (I read English medical journals and technical documentation, write emails and Slack messages), but my speaking and listening aren't strong. Eight years ago, this was a dealbreaker. In 2026, I think this problem is widely misunderstood.

Two reasons. First, AI is currently the most language-capable entity on the planet — Claude, GPT, Gemini are all trained on language at their core. Language proficiency has shifted from "a human skill" to "AI's home turf." Reading English docs, writing English emails, participating in English Slack threads — all of these can now run through real-time AI translation at production quality. The only remaining weak spot is "live English voice meetings," and that use case's share of communication is shrinking fast. Second, native Chinese still gives a context-level edge in AI collaboration — not the often-cited "Chinese saves tokens" claim (our six-task benchmark actually shows Chinese consumes 10–60% MORE tokens than English), but character density still pays off at the context-window level: 1M tokens fits roughly 850K Chinese characters, enough for 3–4 full books. Heavy workflows like "drop an entire contract, repo, or manual into one conversation" favor Chinese users. Combined with my ability to work across Traditional / Simplified Chinese and Japanese markets, this is a more complete toolkit for APAC roles than native English alone.

The old rule — "you need fluent English to work at a global company" — belongs to the last era. The new rule is how efficiently you collaborate with AI + how deeply you can solve problems. Under the new rule, my combination of "native Chinese + solid English reading/writing + heavy AI collaboration" is actually a robust configuration — especially for roles that require working across APAC, Japan, and the Greater China region.

I am the living proof of this thesis — if you're looking for this piece of the puzzle, I might be it.