It was October 2012, and I had just quit my job teaching English at Aeon.
I'd tried to carve out a different path inside the company. I built a multilingual student database on my own time and dime, even presented it to the company president. But the message came back clear: for foreigners, there was exactly one career trajectory beyond classroom teacher. Training other teachers. That was it.
I wanted more.
The timing, of course, was terrible.
I was engaged. Not just engaged, but already the black sheep. An American marrying into a Japanese family. Not just any family, they were all highly accomplished doctors. And here I was: dead broke, unemployed, chasing some vague dream of building software in a country where "startup founder" wasn't a respectable title. It was barely a concept.
When I had asked her father permission to propose to Mari, her father's response wasn't congratulations. It was two words:
"Try harder." [1]
I made a promise to my fiancée that October. Give me until January. Three months. If BiFluent, my little language-learning startup, wasn't showing any traction by then, I'd go find a real job.
She wasn't happy about it. But I suppose a timeline was better than no timeline, especially with a wedding scheduled for April.
January came. And I had nothing to show.
I wanted BiFluent to work. I really did. But I couldn't find an honest path forward. No users. No revenue. No signal that we were onto something. So I started looking for work.
It was dispiriting in a way I wasn't prepared for. The fall from investment banker, to ESL teacher, to can't even get a job as an ESL teacher. I applied everywhere. Finally, in March, I landed a position at a nearby immersion preschool. Teaching two- and three-year-olds.
It paid the bills. Barely.
But here's the thing about teaching jobs: there's a lot of downtime. And my brain couldn't stand still.
I carried a notebook everywhere. Ideas. Sketches. Half-baked product concepts. During nap time, I'd scribble. After work, I'd code. Nights and weekends, I kept building. I kept sharpening my skills, refusing to let go of the dream that someday I'd do something real in software.
My friend Patrick was a translator in Tokyo. We'd worked together at Aeon. I was wildly ignorant of what professional translation actually involved, but one thing I noticed watching him work: it seemed inefficient. He was in Microsoft Word, translating documents line by line, with no way to benefit from repetitions. If he'd translated the same phrase last week, he had to translate it again from scratch.
I did some research and discovered something called CAT tools—Computer-Assisted Translation software. I asked Patrick if he used them.
"Too complicated," he said. "Too expensive. The learning curve is brutal."
Well, I thought, that sounds like a fun project.
What if I could build something just for him? Not as powerful as the big enterprise tools, but simple. Something that could at least give him a small efficiency boost without requiring a PhD to figure out.
So I started building transdraft.
In hindsight, my ignorance was laughable. I was rediscovering concepts that anyone in the industry already knew. The database tables I created didn't even use the right terminology. But the vision I had for simplicity—that seemed to resonate when I showed it to people. After a few iterations, I decided to see if I could actually market this thing.
That's when I found ProZ.com.
ProZ.com was the largest community of translators in the world. They had ads on their site, so I called up their sales rep, a guy named Drew, to ask about promoting transdraft.
Drew, bless his heart, could have just blown me off. But being the kind of guy he was, he took twenty minutes to actually look at my site.
Then he pointed out ten or fifteen things that were missing. Features translators would expect. Terminology I'd gotten wrong. Workflows that didn't make sense.
As kindly as he could, he told me: "I'm not going to take your money. This just isn't ready."
I loved that feedback.
Over the next twenty-four to forty-eight hours—I can't remember exactly—I cranked. I added everything he'd mentioned. Then I called Drew back.
"Hey Drew, could you take another look? I added all the stuff you told me about."
"Get out of here," he said. "There's no way."
"No, really. I added them."
He looked. Paused. "I've never seen anyone move that fast."
Then he offered something that would change the course of my life.
"Why don't I introduce you to our president and founder, Henry? You both have a Japan connection. Maybe he can give you some feedback."
It turned out there was one specific thing Henry really liked about transdraft.
In the header, I'd added a link that said "Find Translators." It didn't actually do anything. It was more of a placeholder, a half-formed idea. But Henry was intrigued that I'd thought to include that angle in a CAT tool. It was something he'd been wanting to pursue for ProZ.com.
So he made an offer: he'd "buy" transdraft and bring me on to work for ProZ.com.
Now, let's be honest—transdraft had no customers and no revenue. The "acquisition" was really just a signing bonus dressed up in fancier language. But it was more than fair by any measure. And more importantly, it was something I'd been chasing for years: A remote software job.
I think about that sequence of events more than I probably should.
What if Drew had just taken my money and let me run a doomed ad campaign? What if he hadn't spent twenty minutes giving honest feedback to some random guy with a half-baked product? What if I'd waited a week to make those changes instead of pulling an all-nighter?
The whole thing hinged on a series of small moments. A sales rep who cared enough to be honest. A founder who noticed a link that didn't even work. A promise to my fiancée that forced me to stop pretending and start executing.
My father-in-law had told me to try harder. I don't think this is what he had in mind.
But it worked out.
That was 2013. A decade later, the lessons from that period still echo in how I build products.
Move fast when you get real feedback. Don't defend your work, improve it. Find the people who care enough to tell you the truth. And never underestimate what can happen when you're too naive to know something is impossible.
Sometimes ignorance isn't a liability. Sometimes it's the only thing that lets you keep going.
[1] It's not as bad as it sounds. He's the most lovable guy, it was really just a mistranslation of the Japanese 頑張ってね ("do your best").