news

in august 2023, shortly after completing their sophomore years at georgetown university and harvard university, brendan foudy（Brendan Foody）with friend adarsh ​​shiremat（Adarsh Hiremath）and surya medha（Surya Midha）they met and then began to persuade them to drop out of school.

in january of that year, the three former high school classmates started a side business called mercor, aiming to connect indian engineers with startups in need of freelance programming help. in just a few months, they hit $1 million in revenue and netted $80,000 without any outside funding.

as soon as fudi’s lobbying session ended, meda turned to shiremat, his partner who won the national championship in policy debate two years ago, and asked a question: “man, how hard is this? "

for the three founders of mercor, the answer seems to be "not difficult."

their software uses artificial intelligence to vet and interview candidates and then match them with open positions. in less than two years, they have conducted more than 100,000 interviews and evaluated 300,000 people. today, at just 21 years old, fudi, shiremat and meda have all dropped out of school and become thiel scholarship recipients, running a profitable company with tens of millions of dollars in annual revenue and 50% monthly growth.

the 20-month-old startup is now valued at $250 million, following benchmark’s $32 million series a funding round, said benchmark’s new partner victor lazarte（Victor Lazarte）also joined mercor's board of directors. billionaire investor peter thiel, twitter co-founder jack dorsey（Jack Dorsey）, two openai board members, and adam angelou, ceo of q&a website quora（Adam D' Angelo）and former u.s. treasury secretary lawrence summers（Larry Summers）they have also joined the ranks of mercor investors.

"we're interested in the labor market, especially when we think about talented people around the world who are not getting the opportunities they deserve," explains mercor ceo fudi. "by applying large language models and letting machines operate with talent assessment with relevant processes, we can address this inefficiency.”

originally a service that manually matched and placed talent, mercor now relies on its own large-scale language model, which is built on the latest off-the-shelf solutions from companies such as openai and based on its expertise in the job search process. the data has been fine-tuned.

to participate in an interview, applicants need to upload a resume and then undergo a 20-minute video interview moderated by mercor's artificial intelligence, half of which discusses the candidate's experience and the other half of which conducts relevant case analysis. after that, the job seeker's application will be matched with all positions on the mercor platform; for positions that require more professionalism, a second round of customized artificial intelligence interviews may be conducted.

for employers, mercor promises to quickly match qualified candidates and be able to flexibly adjust the type of candidates (hourly, part-time or full-time jobs).

mercor's largest candidate pool is in india, followed by the united states. job categories posted also vary and include areas such as engineering, product development, design, operations, and content creation. fudi said some of the top artificial intelligence labs (whom he could not name) use mercor to find professional writers and video producers to create more data. a number of employers, including companies in the consulting, engineering, financial and legal fields, are also looking for employees here. before february this year, mercor itself relied entirely on talent on these platforms, and only recently recruited 15 permanent employees.

according to fordie, mercor's ability to predict candidate performance has exceeded that of human recruiters in the company's tests. he said the company will publish a paper in november sharing its methodology. mercor also hopes to eventually predict and automate long-term performance reviews to help employers track and develop talent. "there is already huge economic value in surpassing humans in predicting human capabilities...but we believe that adding the long-term tracking component will be more sustainable and generate greater economic value," foudy said.

mercor is not fudi and shiremat’s first venture. bellarmine preparatory school in san jose before dropping out of georgetown university（Bellarmine College Preparatory）while studying, fudi ran a small business and built websites for start-up companies. shiremat, who was preparing to enter harvard university, met fudi through participating in the school's debate team and helped him with coding; shiremat's friend mida later became roommates with fudi and joined the entrepreneurial project.

when participating in debate competitions, the founders spend months researching various positions before debating them pro and con. as the children of immigrants from india, shiremat and meda are particularly concerned about discussions of the labor market, particularly employment inequities in developing countries, that prompted them to hire engineering contractors remotely in india. mercor company was established with help. “it feels deeply unfair that the recruitment and screening process arbitrarily eliminates half of the applicants, often simply because of their resume, lack of experience at a prestigious school, or insufficient educational background,” mida said. “building a fairer recruitment method is the best solution to improve people’s overall quality of life.”

last fall, after working in new york for a month, mercor's three founders moved to san francisco to start their entrepreneurial plans full-time. as ceo, fudi focuses on product and sales; as coo, mida oversees operations and enterprise contracts; and as cto, shiremat oversees engineering and machine learning functions. in january this year, they announced a $3.6 million seed round led by general catalyst; in march, all three received the thiel scholarship, which provided them with a $100,000 bonus, exposure to other founders and technology executives, and become friends with them and even get investment opportunities.

mercor's profit margins were growing, so when benchmark's lazarte met with fordie in 2024, the latter had no real interest in raising capital. it was lazart’s own experience in starting a business that made fudi accept the meeting; and investor peter fenton, who is listed on forbes’ list of the world’s best venture capitalists（Peter Fenton）the offer to fly fudi to a helicopter also earned benchmark another meeting. semi-retired former uber investor bill curley（Bill Gurley）later, they also met with the founding team of mercor, which enabled benchmark to fight off other venture capital companies and facilitate mercor’s financing round in advance. (before that, the co-founders had had discussions with vcs more than once because they were dissatisfied with the valuations proposed by them.)

"in the venture capital industry, we meet with a lot of people, but every once in a while you get one that makes you think, 'wow, this person is really cool,'" lazarte said. "i think fordie is a very special person, and this company is about helping people find the right job, which i think is a very noble vision."

but the idea would face a tougher test if companies used mercor to cut jobs and rely on outsourced talent, or to help ai research labs train models to replace other experts. lazarte argued that benchmark wanted mercor to create better jobs for people, not accelerate the process of replacing jobs with artificial intelligence. mercor will also solve the more complex puzzle of matching workers to full-time salaried jobs over time, he added.

mercor faces some well-capitalized competitors in the talent market, such as startup unicorn andela, which recently hired an uber executive as its new ceo. traditional outsourcing companies may also compete using off-the-shelf ai tools. another challenge the company faces is how to keep employers and job seekers in the market interested. fordie admits mercor conducted linkedin searches for certain job postings（LinkedIn）targeted promotion, but he insists that most new users join spontaneously through resources such as resume feedback and interview exercises provided on mercor’s website.

for mercor, a sign of success is "creating a unified global labor market" that understands what everyone's ideal job is and offers it to them. mida summed it up more simply: "if we can help a billion people find jobs, then i'll be very satisfied."

this article is translated from:

text: alex konrad

translation: vivian