The first piece of software Prasanna Sankar remembers with any fondness was closer to a practical joke than a product. A school friend wanted a copy of a game called Magic Claw. Sankar supplied the CD, plus a small script that changed the Windows boot screen to a picture of Tamil film star Rajinikanth and imposed a five-minute countdown. An email address offered the supposed antidote. The next day, the friend arrived furious. Better still, from the young programmer's point of view, the friend's father had written to ask for the antivirus.
It was juvenile and effective. More important, it was immediate. A few instructions written in one room had rearranged the atmosphere in another. That feeling - the private logic becoming a public consequence - has followed Sankar from coding contests in India to Microsoft, from the brief blaze of a college social network to the long construction of Rippling, and now to an AI company called Vorflux.
His career is often summarized by the largest number attached to it. Rippling, which he co-founded with Parker Conrad in 2016, was valued by private investors at $16.8 billion in May 2025. Forbes estimates that Sankar still owns about nine percent. Yet the more revealing numbers may be smaller: the three Y Combinator rejections he says came before acceptance, the roughly two years Rippling spent building in stealth, and the six weeks in which his first company acquired 20 million page views and learned that velocity can be a splendid disguise for fragility.
The student who preferred the scoreboard
Sankar grew up in Chennai and has described himself as an unremarkable school student. The family computer first offered games, then something more agreeable than the classroom: a system whose basic operations could be combined into almost anything. He liked that expertise could be challenged. Start with the instructions, reason forward, see whether the machine agreed.
At the National Institute of Technology in Tiruchirappalli, that instinct found a scoreboard. Sankar ranked first in India on Topcoder while in college and represented NIT Trichy at the ACM-ICPC World Finals in 2007 and 2008. The surviving contest record is pleasingly unsentimental - 71st one year, 80th the next. He reached the Google Code Jam final too. The achievement was not a parade of gold medals. It was proof that he could train his way into rooms that had once seemed inaccessible.
“You can build everything from first principles.”Prasanna Sankar on why coding caught him
An internship at Google followed in 2006. After graduating in 2008, he chose Microsoft over a Google offer for the disarmingly practical reason that Microsoft paid more. On the Bing team, he worked in low-level C and C++ across large fleets of machines. Early on, after an evening request exposed a missing use case, he stayed up through the night and sent a patch the following morning. It crossed several components of a codebase he had barely met.
The place should have suited him. Bing was growing quickly and the problems were substantial. But patches moved slowly, the last fraction of the work stretched on, and Sankar began to fear that being India's top-ranked programmer would remain the most distinctive sentence in his biography. He stopped going to the office. Four months passed before Microsoft fired him. It is an almost comic exit from respectable employment, except that Sankar took the underlying dread seriously: what if his peak was already behind him?
Twenty million reasons to be wrong
Silicon Valley supplied a more dramatic exam. Sankar and his co-founders tried several ideas before a phone call from his then-girlfriend in Cambridge redirected them. A student social network called FitFinder was spreading there. She suggested adapting the premise for Stanford. They did. LikeALittle let students post anonymous flirtations tied to campus locations - a line about the person in green jeans at the library, perhaps, followed by a crowd trying to identify both parties.
The product spread with the efficiency of gossip. About a fifth of Stanford used it in the first week, Sankar has said. Within six weeks, it had recorded 20 million page views across the United States. Andreessen Horowitz, Ashton Kutcher and Yuri Milner invested. The company joined Y Combinator after Sankar's team had already been rejected several times. LikeALittle acquired the usual ornaments of arrival before it had acquired permanence.
Retention weakened on the early campuses. The team attempted a move to mobile, but many college students did not yet own iPhones, and desktop users did not travel neatly into an app. Later products accumulated millions of downloads and then emptied out. Sankar discovered a problem competitive programming had not prepared him for: people are not test cases. They do not remain inside the system merely because the system works.
“The biggest part of the burnout for a failed founder is actually the pretending.”On maintaining confidence while a company fades
He has spoken bluntly about the performance required during failure - dinner with successful founder friends, a brave face for employees, the fear that candor itself could hasten the end. When the pivots were exhausted, he left the Valley, moved to the Netherlands and expected to stop founding companies. LikeALittle had given him the intoxication of growth and the hangover of an audience that could leave without explanation.
A system that people could not simply forget
The opposite of a fickle campus audience was an enterprise customer with payroll to run. From the Netherlands, Sankar searched for a young company growing quickly and found Zenefits. He joined as director of engineering and became fascinated by its founder, Parker Conrad, whose strategic predictions Sankar had watched survive contact with the market. The two argued about pricing and product late at night. When Conrad left Zenefits in 2016, Sankar emailed to say that he wanted to work with him again.
Rippling started in Conrad's San Francisco home. The ambition was inconveniently broad: one employee record connected to payroll, benefits, devices and software access. An ordinary startup could ship one small wedge and learn. This one required a wide foundation before the demonstration made sense. Sankar and Conrad spent roughly two years in stealth. By late 2017, 14 people were working around Conrad's family and dog.
The product eventually clicked. Thousands of customers arrived. Sankar led an engineering organization of about 100 people, but the work changed as the company did. His days filled with meetings designed to keep teams aligned. He has called himself ineffective in that role, not because alignment was unimportant but because he did not enjoy doing enough of it. Stepping down as CTO in 2020 was a rare piece of founder self-editing: the company needed someone energized by the job that now existed, not the one he had loved four years earlier.
The same irritation, in new clothes
Sankar did not lose interest in connected systems. With 0xPPL, founded in 2023, he applied the idea to crypto activity scattered across wallets, chains and smart contracts. The 2.0 version announced in 2025 combined a wallet, portfolio tracker, bridge and app interface. Its premise resembled Rippling's at a distant angle: the user should see one coherent surface where the underlying world offers a box of unrelated parts.
Then AI coding models improved, and Sankar found another box of parts. In July 2026 he launched Vorflux with co-founders Anant Nag and Haritha Elango and announced $15 million from Y Combinator, Peak XV Partners, Powerset, Alliance, Parker Conrad, Immad Akhund and Balaji Srinivasan. The company does not train its own foundation model. It builds the harness around existing ones: an environment that plans work, writes across a codebase, runs tests, reviews changes and prepares a production-ready pull request.
Sankar's argument is that the bottleneck has moved. Code generation is increasingly cheap; human supervision, environment setup, validation and coordination remain expensive. “Human time is the bottleneck. Tokens are not,” he wrote at launch. It is a bold product thesis, still young enough to deserve scrutiny rather than coronation. Funding is evidence that accomplished people find the question worth asking. It is not an answer.
What gives Vorflux its biographical neatness is the way it answers Sankar's old complaints. At Microsoft, the patch came quickly and the organization moved slowly. At Rippling, the programmer became a manager of calendars. Now he is building software meant to keep moving after the programmer leaves the keyboard. The boy with the prank CD wanted code to produce a consequence. The adult founder wants the consequence without an afternoon of status meetings.
The industries in his career look restless when listed together: campus social, human resources, crypto, artificial intelligence. The underlying irritation is remarkably stable. Sankar notices a landscape broken into too many steps, then tries to compress it into one system. Sometimes the crowd arrives and disappears. Sometimes the system becomes a multibillion-dollar company. Either way, the next blank screen appears, and he sits down in front of it.