Have you hired software developer Soham Parekh yet?

Mint

a day ago

A few days ago, Soham Parekh was just another full-stack developer.
Then his name began surfacing across founder circles: Hacker News threads, Slack channels, Twitter jokes, Reddit threads. One YC-backed startup after another realized they'd hired him. Not in sequence—at the same time.
Some found out after a few weeks. One team said he worked with them for nearly a year. The stories converge on the same arc: stellar interviews, fast onboarding, some early output. Then missed meetings. Odd excuses. Gaps in availability. In one case, Soham turned up for a trial in person, then left halfway through the day, saying he had to meet a lawyer.
He didn't disappear. He just kept showing up somewhere else.
The question isn't how he got away with it. The question is why it was so easy.
Soham Parekh is not the first engineer to work multiple jobs in parallel. In November 2022, Vanity Fair published a piece titled 'Overemployed in Silicon Valley: How Scores of Tech Workers Are Secretly Juggling Multiple Jobs." It told of engineers quietly holding down two, three, even four full-time roles. Some used mouse-jigglers to fake activity. Others ran multiple laptops. One admitted to outsourcing work to Fiverr. A few worked in coordinated Discord communities, sharing tactics.
'I'm not sure if they even know I'm here anymore," one engineer told the reporter. 'All my paychecks are still coming in."
At the time, it read like a side effect of the remote-work boom. A strange consequence of too many laptops and not enough oversight. Soham didn't need any of that infrastructure. He used his real name. Real resume. Showed up on video calls. Wrote code. Left a trail. He just moved through the system cleanly.
What his story shows is how little it takes to get hired—and stay hired.
One startup said he 'crushed the interviews." Another called him 'top 0.1%." Founders praised his GitHub, his side projects, his email follow-ups. They only saw the red flags once the real work began. That gap—between performance in a vetting process and actual engagement—isn't incidental. It's structural.
Startups, especially ones chasing growth, have narrowed hiring into structured calls and take-home tasks. Processes are recycled across founder networks. Culture fit becomes a checkbox. Most of the time, it comes down to gut feel. Which is just another way of saying: we don't really know. In that kind of system, someone who interviews well and ships enough can coast for months. If that person is also working three other jobs, the signs fade gradually. By the time someone notices, it's already awkward to ask.
There's another wrinkle. Soham may not have been doing anything that couldn't be done today by an AI agent.
More than one founder joked—what if he was a bot?
That question no longer lands as satire. Agents today can write code, answer support tickets, even joke in Slack. At some point, you stop noticing the difference.
And if you can't tell whether you're working with a disengaged employee or a competent script—what exactly are you hiring?
We've seen this fragility before.
In 2022, Wipro fired 300 employees for 'moonlighting." Chairman Rishad Premji called it 'cheating—plain and simple." The company said some were working for competitors. The backlash was swift. Critics pointed out that many executives sit on multiple boards. Others questioned the demand for loyalty from a system that rarely offers the same in return.
That episode surfaced a buried truth: the rules of work have changed. Expectations haven't.
Soham Parekh is a consequence of that mismatch. He's not a rogue actor. He's the product of a hiring culture that values performance over presence, delivery over connection. A culture that claims to build teams but rarely asks who's actually part of them.
So what happens when the next Soham is indistinguishable from an AI agent?
Srikanth Nadhamuni, the former CTO of Aadhaar, believes we'll need to rethink identity itself. In a recent paper, he proposed Personhood Credentials—a cryptographic and biometric framework to prove that a person behind a digital interaction is real, unique, and singular.
The concept sounds abstract, even dystopian. But Nadhamuni argues that in a world of deepfakes and synthetic voice agents, systems like Aadhaar—originally built for public verification—could help anchor digital interactions to actual humans. He describes it as a privacy-preserving firewall against the collapse of trust online.
It raises real questions. About privacy, about exclusion, about the kind of infrastructure we're willing to accept in the name of certainty. But it also names the thing most companies pretend not to see: if you don't know who's on the other side of the screen, you're not hiring a person. You're hiring a pattern.
And if Soham Parekh passed every test and still wasn't who we thought he was—what happens when the next Soham isn't even human?
Pankaj Mishra is a journalist and co-founder of FactorDaily.