Adarsh Hiremath

Hiremath and his two co-founders cut their own salaries to $500 a month, and he has said it was the pay cut, not the seed round of more than $3 million that followed it, that was the moment they felt they had "made it." He was 19.¹ He is the co-founder and chief technology officer of Mercor, a company that uses AI to evaluate job candidates and matches vetted human experts, engineers, lawyers, doctors, bankers, and other specialists, with companies and AI labs that need their work. Mercor's largest business supplies the human-generated data frontier AI labs use to train their models. Hiremath is one of three co-founders, alongside chief executive Brendan Foody and board chairman Surya Midha, who built the company from a college dorm room into a business valued in the billions of dollars within a few years.

Background

Hiremath was raised in California's Bay Area, the child of parents who immigrated from India. He attended Bellarmine College Preparatory, a Catholic high school in San Jose, and joined its competitive policy debate team. There he met Surya Midha, whom he had actually first encountered around age 10 as one of a small number of elementary-school-aged competitors on the high-school debate circuit, and Brendan Foody, a classmate at Bellarmine. Hiremath and Midha became debate partners, competing nationally; Hiremath has described that partnership as effectively his first startup, built on shared ownership of outcomes and a constant feedback loop. The three later formed the core of a program that won all three major national policy debate championships in the same year.¹

Hiremath enrolled at Harvard University in a combined bachelor's and master's computer science program. He has said he was at Harvard "physically, not mentally," already consumed by the early version of the business that would become Mercor, and dropped out during his sophomore year to work on it full time.¹

Getting started

Hiremath, Foody, and Midha started with no grand ambition: a small dev shop recruiting exceptional software engineers in India and staffing them onto projects for startups, initially for the founders' own friends. Hiremath has said the team came to realize the business was "more about the people than the software."¹ They began using AI to vet candidates, automating first the candidate-facing side and then the company-facing side; that automation became the product itself, and the marketplace that would become Mercor was born. General Catalyst led the seed round, and all three founders later received Thiel Fellowships. They then recognized a broader shift: AI labs were beginning to hire thousands of people to generate the human data needed to train next-generation models, a demand Mercor's infrastructure was suited to serve directly.

What he built

Mercor was founded in 2023. Its core product uses AI to predict how well a candidate will perform in a role, compressing resume review, interview, and hire-decision into an automated pipeline, including an AI interviewer configurable for a new role in seconds. Its larger, faster-growing business recruits and vets specialized experts, engineers, finance and consulting professionals, lawyers, doctors, and scientists, and places them with AI labs and enterprises that need expert-generated data and evaluation sets to train frontier models. Clients have included most leading US AI labs and traditional enterprises.

Growth has been unusually fast: Mercor reached roughly one billion dollars in annualized revenue within about twenty months of a much smaller run rate, a pace the founders call the fastest they are aware of.² The company raised a Series A led by Benchmark and a Series B at a valuation of about $2 billion led by Felicis, with General Catalyst and Benchmark also participating. By October 2025, investors had valued Mercor at roughly $10 billion, and reporting in mid-2026 indicated talks around $20 billion, a valuation that would put the three founders, all in their early twenties, among the youngest self-made billionaires reported anywhere.² Mercor also built its own benchmark, APEX, with Cognition, measuring AI performance on economically valuable professional work such as software engineering, law, medicine, and investment banking, distinct from academic benchmarks like Olympiad math; several major labs have adopted it.³

How he operates

Hiremath's central framing of Mercor's business is that human data has become fundamentally a talent-assessment problem rather than a crowdsourcing one. He believes improving an AI model's performance in a domain requires an expert better than the model at that task, and that identifying that expert is a vetting problem, not a labeling problem. He argues evaluation sets, by definition, must be built by humans who exceed the model's current capability, making high-quality human data the binding constraint on AI progress.¹ He also believes that as software costs approach zero, durable businesses will instead be built on network effects, citing marketplaces like Uber and Airbnb as the model Mercor is chasing on both sides of hiring. He has said Mercor selects for people who "care deeply about the mission" rather than mandating hours, and does not believe intensive schedules work well remotely, part of why Mercor operates in person in San Francisco.¹

Where things stand

As of the mid-2020s, Foody serves as chief executive focused on AI labs and frontier customers, Midha as board chairman, and Hiremath as chief technology officer concentrating on enterprise expansion. The company describes itself as capacity-constrained rather than demand-constrained, and has stated an ambition to become the connective infrastructure for a much larger share of global hiring and AI training data.³

Key facts

• Co-founded Mercor in 2023 with Brendan Foody and Surya Midha, all three from the same Bellarmine College Preparatory debate team, after an early venture as a dev shop staffing engineers from India.
• Dropped out of Harvard's combined computer science bachelor's-master's program to build Mercor full time, and received a Thiel Fellowship alongside his co-founders.
• Cut his own salary to $500 a month before Mercor closed a seed round of more than $3 million from General Catalyst at age 19.
• Serves as chief technology officer, leading enterprise expansion while Foody focuses on AI-lab and frontier customers.
• Reported to be among the world's youngest self-made billionaires after investors valued Mercor at roughly $10 billion in October 2025, with talks reported in mid-2026 around $20 billion.
• Frames human data as a talent-assessment problem, arguing that evaluating AI model performance in a domain requires a human expert better than the model itself.