At three in the morning, John Kim sat inside a booth at the University of Sydney library, conducting business with San Francisco. He was no longer a student there. The campus was simply the available office, and the nocturnal schedule was the tax for building an American company from Australia. Wake at 11 p.m. Work through the night. Sleep after lunch. Repeat until either the company or the calendar changed its mind.
The company was Paraform, a recruiting marketplace Kim had started with Jeffrey Li. It was young enough to be fragile and old enough to have disappointed a real customer. Hightouch had trusted the pair with a director of engineering search, a role that had already resisted agencies, tools and job boards. Then came the message: the candidate quality was not good enough. They had ten days to turn it around.
There are glamorous founder crises, at least after the lighting department has finished with them. This one involved a borrowed library booth, a difficult immigration process and the unpleasant knowledge that a customer was perfectly entitled to leave. Kim kept working. Paraform filled the role in about six weeks. Hightouch said it had been open for 18 months.
The result did more than preserve the account. A customer introduced Kim and Li to investors. The founders bicycled over the next morning without a deck because they had not understood the meeting to be a fundraise. Drinks at five became a term sheet that day. Venture capital rarely looks so much like an errand.
“If you think you’re going deep, you’re probably still not deep enough.”John Kim on staying with a difficult problem
Four countries and one recurring question
Kim grew up moving because of his father’s work. From the age of four, home meant a sequence: Singapore, Australia, Los Angeles and Korea. His grandfather hoped he would study law, and Kim arrived at the University of Sydney with enough intellectual restlessness to appear in both the robotics club and a torts moot. His public record also includes an academic merit scholarship, an engineering education he did not finish and an internship at Tesla.
Recruiting, however, became the durable theme. Before Paraform, he co-founded RulerNetwork, a service connecting students with mentors and career guidance. Crimson Education acquired it, and Kim became a director of product working on careers and recruiting. Many founders claim their first company taught them everything. In Kim’s case, the more credible reading is that it taught him which question he was willing to ask twice.
He met Li on a business trip to New York. They asked founders for their biggest problem and what solving it might be worth. Hiring kept winning; one successful placement could command roughly $40,000. Their first plan possessed the fatal charm of simple arithmetic: make five hires each and they would have something resembling a salary.
They became recruiters. They gathered public signals from X, GitHub, LinkedIn and résumés, then tried to identify the right engineers. Three or four months later, they had made none of their ten intended hires. Data could locate a promising person. It could not make that person reply.
The useful users
The founders changed the product into a referral marketplace. Companies could post bounties, and anyone with a promising connection could collect one after a hire. The proposition was democratic, cheerful and not especially sticky. Most people know fewer job-ready specialists than they imagine. Recruiting was not their profession, so they had little reason to return once their small circle had been exhausted.
One group did return: independent recruiters, agencies and headhunters. The obviousness arrived late, as obviousness tends to do. These users had broad networks, practiced judgment and a financial reason to keep matching people with roles. Paraform stopped trying to turn everyone into a recruiter and began helping recruiters behave like independent businesses without forcing each one to assemble a back office.
The marketplace gives companies access to specialists and gives recruiters roles, workflow software and performance feedback. AI assists with sourcing, transcription, candidate matching and the administrative shuffle. The recruiter keeps the task that determines whether the machinery matters: understanding a person well enough to make a credible introduction.
“Data is commoditized and software is commoditized. But human talent is still the most valuable resource in the world.”John Kim after Paraform’s Series B
Kim’s argument is sharpened by a paradox. Generative AI has made outreach easier, but easy outreach creates a swamp of competent-sounding messages. He has relayed a recruiter’s estimate that one engineering hire took contact with 500 to 700 candidates five years ago, 1,500 two years ago and roughly 5,000 more recently. Those are reported operating observations, not an industry census, but the direction is recognizable to anyone with an inbox.
“Meaning comes from scarcity,” Kim says. A message from a stranger can be generated in seconds. A recommendation from someone who knows your work has a history attached. Paraform’s bet is that software should enlarge the reach of that history without pretending to own it.
The economic design matters because a marketplace gets the behavior it pays for. In 2023, Kim contrasted the few thousand dollars a recruiter might receive for a placement inside a traditional firm with the five-figure average he said recruiters could earn per hire on Paraform at the time. The platform also paid recruiters when a candidate advanced beyond a first interview, giving useful work some value before the all-or-nothing placement fee. Terms can evolve as a company grows, but the original signal was unambiguous: bring skilled recruiters onto the network by making the network good for skilled recruiters.
The backers around Kim and Li make a second network visible. Early money included Primer Sazze Partners, Twitch co-founder Kevin Lin and leaders connected to Scale AI and Founders, Inc. Later rounds brought in A*, BOND, DST Global, Liquid 2 and Felicis, followed by Scale Venture Partners. The cap table has also included founders or executives associated with Canva, Instacart, YouTube and other technology companies. Those names are less interesting as decoration than as distribution: people who have built companies know many more people trying to build them, and every growing company eventually confronts the same awkward question of whom to hire next.
Kim’s public writing increasingly treats the individual hire as the atomic unit of company history. There is, he argues, a version of every business in which one more person arrived: the builder who unlocked a product, held a team together or opened a new stage of growth. It is a founder’s view of recruiting, sentimental at the edges but severe in its consequences. Payroll is a spreadsheet; a team is a series of bets on particular people.
Changing the mechanism, keeping the problem
Paraform’s financing now forms a brisk little staircase: a $1.4 million pre-seed in March 2023, a $3.6 million seed in April 2024, a $20 million Series A led by Felicis in June 2025 and a $40 million Series B led by Scale Venture Partners in March 2026. Total capital reached $65 million. Forbes placed Kim and Li on its 30 Under 30 AI list for 2026.
Kim and Li found Paraform and begin testing recruiting models.
The company announces $1.4 million in pre-seed backing.
A* leads a $3.6 million seed round.
Felicis leads the $20 million Series A.
Scale Venture Partners leads a $40 million Series B.
The customer list moved beyond early startups to organizations including Palantir, Rippling, Cursor and Abridge. The platform moved beyond engineering and go-to-market searches, too. By the Series B announcement, Kim said it was already working with about 30 law firms and had not needed to perform major surgery on the product. A trustworthy broker, it turns out, travels rather well between professions.
Kim and Li’s May 2026 manifesto widened the ambition again. They described Paraform as a talent company built around hiring intelligence: use the outcomes of searches and interviews to understand fit, then turn those signals into agents that can run more of the workflow. Kim has said the company grew revenue tenfold in 2025 and paid $50 million to recruiters by May 2026. In August, he publicly said Paraform had reached $100 million in annualized revenue.
The more durable detail is how Kim describes the work. His advice is not to cling to the first product. Paraform discarded several. It is to maintain conviction in a consequential problem and keep asking why the current answer fails. The first attempt treated hiring as discovery. The next treated it as casual referrals. The version that worked treated it as a profession held together by relationships, incentives and repeated judgment.
That interpretation also explains Kim’s temperament. He speaks plainly about failure, gives specific numbers when recounting an experiment and returns often to responsibility. During the library stretch, Paraform found customers without automated email: knock on doors, write to people individually, regard each account as precious. The chief executive of an AI company appears to have learned his favorite lesson from manual labor.
Years from now, Kim has said, the present may look like that 3 a.m. booth looks today. A difficult period becomes a small room in the memory, its fluorescent light softened by everything that happened afterward. His imagined advice to the younger founder is spare: keep going, have conviction.
It works because the conviction is not mystical. Paraform’s story contains a practical sequence available to anyone building a marketplace. Do the job yourself. Notice which users return. Separate abundant information from scarce trust. Automate the surrounding friction. Pay the person who supplies the judgment. Then, if an investor calls, keep the bicycle nearby.