In seven years I started five companies. I made money, lost money, and owed six or seven million. I started Brightstream because I realised AI is the biggest lever this generation of owner-operators will get — and I wanted to share that road with more owners like me.
The nickname has a simple origin: I am built like a bear and I am steady. Over time friends and clients just started calling me that —steady and reliable, in character and in build, like a bear.I like it, because it happens to name exactly what I want to be remembered for: reliable.
Finished my master's early; the youngest senior engineer in the whole R&D building, with a manager 13 years older
Walked away from a seven-figure salary. Dr. Chuang HairScience passed NT$1 million in course sales in month one
Ran it for 3 years, past NT$10 million in sales
4 years; learned cash-flow discipline, that a shareholder is not free labour, and how to design a profit share that does not backfire
Turning every bit of tuition into a new company
A one-person core, fully remote, past NT$10 million in revenue in 10 months
A founder's real CV is not a list of wins, it is a list of failures. Here are three I am willing to lay out — because this is what you will actually hear across the table if you come to work with me.
I thought taking a business already doing eight figures in Taiwan into a market ten times the size was a scaling problem. It was a rebuild: book numbers, licences, a pandemic that would not end, and a local marketing agency that walked us through a pile of useless work. That chapter cost several million.
A bigger market does not mean a bigger opportunity; if you have not understood the rules, a bigger market kills you faster.
4 years as co-founder and CTO. It was profitable at first on the core systems business, then died because fixed costs rose too early, the product shipped with no money left to promote it, too many side quests opened, and cash flow would not hold. Years on a low salary, covering company costs out of pocket.
A company does not survive on ideas; if cash flow will not hold, even a good product is a side quest. Being a shareholder does not mean working for the company for free.
Brightstream is not a company I started on a good run. Before it existed, the debt stacked up from the earlier ventures, plus what my family had put in, was all on me. Brightstream is the restart I bought with six or seven million in tuition.
So when I say a small company cannot afford to waste money, it carries weight. I am not teaching swimming from the shore.
I have started five companies across seven industries (that is my own founding history, not Brightstream's client count). When I say "cash flow", "a gap in the team", "the ads will not convert", those are things I paid for personally.
I am not going to explain transformers or RAG to you. I will work out with you whether next quarter earns NT$2 million more, and whether the team still runs when you take a week off — that is what an owner cares about.
Every engagement starts with a free 30-minute diagnostic where both sides assess each other. If we go ahead, I think about your problem the way I think about my own — not as a project.
Naming it Brightstream was putting my own name on the door — I am not selling a product, I am selling judgement.
What we do is simple: help owner-operators use AI to run the business right and make it bigger. We focus on three industries (medical aesthetics, print, D2C) because the custom systems they need are complex and heavy enough to match the depth we can actually deliver.
How many deep engagements can one owner run in a year? The most I have seen is 8–10. So Brightstream will not scale without limit — we choose who we work with, and we see every project through.
「Judgement matters more than tools.」
「Owner to owner, no engineer-speak.」
「Deep is harder than plentiful.」
「Custom systems, not templates.」
「Every engagement starts with a free 30-minute diagnostic where both sides assess each other.」
「Partners on the same side of the table, not a vendor.」
「I do not take work that goes against what I believe.」
「It only counts as success when the business actually changes.」
We are not a software vendor,
and not engineers billing by the hour.
What we want with a client isto grow together and keep doing well together。
Because our definition of a client relationship is "long-term", the cases where we cannot work together have to be said out loud first. Whether a consultancy can be trusted comes down to whether it is **willing to say no** — the clients who stay after that are the right ones. So these five are published here, and we know before the first meeting whether we fit.
"Can you just throw it together in WordPress", "can you copy that other company's site", "do you have a template you can drop on top"
A template is exactly that — one generic layout with your logo and colours on it — and when it ships you look like your competitors. For an owner who is serious about the brand and about a genuinely custom sales process, that is not saving money, it is giving up differentiation. We build 100% custom systems where every component and every step follows your way of working. Plenty of people in this market will do that cheaper than we will.
"You take all of it, I do not want to hear about the process", "here is the data, call me when it is done", "you are the experts, you decide"
Our method runs advisory and engineering in parallel. The judgement calls that actually matter in a custom system — how to price, where the process jams, which step AI belongs in, which features to cut — only the owner is qualified to make. Hand over the data and the responsibility goes with it, and the chain that makes this method work snaps. Custom becomes fake-custom, and what ships is software you do not want and will not use.
"Just build it, we do not need the ongoing work", "launch it after the demo and we will take it from there", "do a cheap version first and see how it goes"
Sales keep moving, the market keeps moving, the models keep improving — the moment a custom system is frozen, it starts losing value. The most valuable thing we have is know-how accumulated by landing and expanding: in three years your system will look nothing like today's, because it grew with your business. A short one-off build leaves you paying for software you cannot maintain, and leaves us unable to deepen vertical know-how. Both sides lose.
Three-way RFPs where lowest wins, "the market rate is X", "that is too expensive, I can get a student to do it", picking a consultant the way you run a procurement tender
Expertise has a price — your own field surely has a reasonable range too, and pushing below it means giving up quality. We are not a vendor, we are a partner, and partners do not enter bidding wars. If your way of choosing a vendor is to get N+1 quotes and take the lowest, then what you want is not a partner, and other people in this market will do it cheaper. We respect that choice — it just is not us.
"I do not know what I want, give me some ideas", "AI is hot right now, we should do something", "let us do a website first and see"
We are here to solve a problem, not to work out what you should be doing. If you have no clear target, goal, problem or pain point — something like "GMV up 30% next quarter", "support response down from 24 hours to 30 minutes", "dropped sales leads under 5%" — the free 30-minute diagnostic will give you advice but we will not take the project. Custom development with no north star produces something you will not use and your team will not use. My advice is to work out the problem first, then come to us.
Do not read this list as posturing. Read it as the most honest opening line I can give any owner who comes to talk —
Whether I take you on has to be settled before whether you take me on.
