Before Enoch Ko could explain himself to his American classmates, he could show them a game. It ran on his father's TI-84 graphing calculator, a gray brick designed for equations and, in the right hands, mischief. Ko had arrived in the United States at 11. He knew little English. The calculator offered a small, glowing stage where verbs were optional. Look, the program could say. I made this.
The classmates were impressed and, by Ko's own amused recollection, probably thought he was a little weird. A reasonable verdict. He kept going. He wrote more games, then a programming language and a game engine for the calculator. His father, a math teacher, occasionally confiscated the machine because seven hours of calculator play did not resemble studying. The family joke has aged well: the forbidden toy became a route into computer science, and computer science eventually became a route into one of American business's most paperwork-heavy rituals.
Today Ko is the founder and CEO of Soraban, software for accounting firms that handles the bustle surrounding a tax return: collecting documents, organizing files, moving data into familiar tax systems, preparing workpapers, securing signatures and nudging clients who have forgotten that one stubborn form. Its territory is not the dramatic calculation at the center. It is the queue outside the calculation, tapping its foot.
Three countries and a talent for starting over
Ko was born in Delhi to Korean parents who were studying and doing missionary work in India. He and his two younger brothers spent their earliest years there. Hindi was his strongest language, with some English and Korean around the edges. At about seven, the family moved to South Korea. The language, food, school and social code changed at once. Just as he found his footing, the family moved again, this time to the United States.
He describes the move at 11 as the hardest. Childhood friendships were beginning to feel durable, then the map was redrawn. For a while, his answers in English were largely limited to yes and no. The experience forced invention. Origami had already supplied one refuge: as a young child, he could fold paper for seven or eight hours a day. Lego supplied another. He invented long-running games for his brothers, complete with rules and statistics written on paper. The TI-84 joined that collection of homemade languages.
Moving repeatedly did not turn rejection into fun. It made rejection survivable. Ko says he learned not to be afraid of looking foolish. A story could fall flat; an invitation could go nowhere; another attempt remained available. He also learned to look for more than one acceptable way through a problem. Those are founder-friendly traits, though in his telling they began as practical childhood equipment.
“I think the biggest thing I've felt and learned is not being afraid to look like a fool.”Enoch Ko, PHX Founders interview
The family's economics supplied another lesson. When five people moved to the United States, his father's teaching salary was the main income. Ko remembers seeing how far a limited amount could stretch when choices were deliberate. Years later, he would connect that memory with Soraban's restrained early spending. Constraint, in his view, can sharpen creativity. Money is useful; attention to what money is for is better.
The tax bill that changed the assignment
Ko studied computer science at Arizona State University, then worked in software engineering and consulting. In 2018, he was juggling four or five clients and earning well. The success arrived with an unpleasant receipt: he recalls a tax burden near $80,000. A friend who was a CPA explained deductions, corporate structure, payroll and QuickBooks. With that help, Ko says he saved more than $40,000.
What bothered him was not simply the bill. It was the possibility that the knowledge was obvious to everyone else. Ko had grown up with a different tax system and wondered whether Americans had taken some hidden Tax 101 class. He asked other contractors. Many were equally confused. So the software engineer did something unusually committed for a curious outsider: he helped start an accounting firm.
A large tax bill and a CPA friend's advice reveal how much tax knowledge independent workers can miss.
Ko operates a technology-enabled accounting firm and encounters the workflow from inside the service.
Soraban enters Y Combinator's Winter batch and turns toward software for accounting firms.
The company announces a Series A led by Altos Ventures and broadens its capacity-focused story.
Prepare extends the platform into workpapers while keeping accountant review and control explicit.
The first version focused on self-employed people and small businesses, including contractors considering an S corporation. Ko partnered with accountants and spent roughly two years operating the service. He built bookkeeping software, connected systems and designed forms. He also discovered that adding technology to a firm did not make the firm's daily friction politely disappear. Clients still had to be chased. Files still arrived in odd places. Staff still waited for missing information. The founder had entered the machine and found loose belts everywhere.
That experience changed Soraban's customer. Rather than remain a technology-enabled firm for taxpayers, it became software for the accounting firms themselves. The distinction matters. Ko did not study the workflow from a survey and prescribe efficiency from a safe altitude. He spent tax seasons inside it, where every tidy diagram eventually meets somebody's upside-down phone photo of a W-2.
A product built around the work nobody frames
Soraban describes the accounting back office as an execution problem. A conventional workflow tool may record that a document is missing or route a task to an administrator. Ko wants the system to perform more of the task: ask the client, receive the file, name and sort it, validate the data, move it into tax software and preserve a trail a reviewer can follow.
The boundary he draws around artificial intelligence is as important as the automation. When Soraban introduced its Prepare workpaper system in June 2026, Ko summarized the compact: “The AI handles the prep work, the accountant makes the decisions.” The system was presented with manual edits protected and AI-generated review items dismissible. Judgment, credentials and accountability remain with the professional. The machine gets the folders.
The company's name makes the same argument with older technology. A soroban is a Japanese abacus, refined into an instrument of speed and precision. It does not pretend that arithmetic no longer requires understanding. It gives trained hands a better interface. Ko's version of accounting AI is similarly concerned with augmentation, though the beads are now documents, reminders and data fields.
The discipline of a weekly number
Soraban entered Y Combinator with Ko as a solo founder and, he says, little more than an idea. The Winter 2021 program was remote, which meant a company rooted in Arizona could join without the gravitational tug of three months in the Bay Area. Ko later recalled raising about $1.3 million over Zoom in a few days after the batch, an experience that weakened the assumption that every software company needed the same postcode.
YC's weekly check-ins were less liberating. Founders reported a key number, commonly revenue, in front of peers. Whatever else had been accomplished, the question returned: had it moved? Ko describes the process as stressful and clarifying. Conferences, founder social life and ornamental activity fell away. The company had to learn what could improve the product and persuade a customer this week.
That focus met the thrift he learned at home. Ko argues that spending and growth are not mechanically linked, and that a little less capital than desired can force a team toward product-market fit. The point is not austerity as theater. His family stretched a teacher's income because it had to. A startup should treat each constraint as a reason to learn faster, not as an excuse to decorate hardship.
By 2025, Soraban announced a Series A led by Altos Ventures and said it was powering more than 300 accounting firms. Its public materials now cite more than 350 firms, six tax seasons and a 50-person team. The product has widened from intake into preparation, tax-software connection and delivery. The ambition on its Y Combinator page is to serve tens of thousands of firms and build for decades. This is patient language attached to a profession that measures time in filing deadlines.
“The AI handles the prep work, the accountant makes the decisions.”Enoch Ko, on Soraban Prepare
The connection hiding inside the workflow
Ko's story has an appealing symmetry. As a child, he made objects that helped him connect when language failed: folded paper, elaborate games, a tiny program shared across a desk. As a founder, he builds software around the failed connections of tax season: the request a client overlooks, the document an administrator cannot locate, the number a preparer cannot trace.
The comparison should not be polished into destiny. A TI-84 game does not forecast a Series A, and a childhood move does not arrive with a product roadmap. What it can create is a habit: enter an unfamiliar system, watch where people get stuck, make something concrete, risk looking foolish, revise.
There is also a quiet joke in Ko's choice of industry. The boy who once worried Americans had secretly mastered taxes now runs a company devoted to the fact that even professionals deserve better machinery around them. He did not discover a missing class. He discovered missing infrastructure.
Soraban's wager is that accounting firms do not need more dashboards observing work. They need some of the work to happen. Ko's longer wager is more personal: that careful tools can translate complexity into action without removing the human being responsible for the answer. It is the calculator lesson once again. Let the instrument move quickly. Let the person decide what the numbers mean.