Mohit Bhende and the value of one more chance

Mohit Bhende applied to 13 colleges. Twelve rejected him. His father supplied a shorter assessment: “all it takes is one.” It is a good line for a young person facing a stack of refusals. It also makes a useful starting point for understanding the company Bhende would eventually co-found. Karat built a business around the technical interview, including the rather humane proposition that a candidate should sometimes be allowed to try again.

Bhende, known as Mo, grew up in India and went on to Penn State, graduating in 2001. The later credentials are formidable: Cambridge, a Gates Cambridge Scholarship, an MBA from Wharton. But the earlier arithmetic deserves to stay in the picture. Twelve decisions about an applicant did not exhaust what that applicant could become. A hiring process has to make judgments; it also has to live with the people those judgments miss.

His work now concerns that awkward interval between having ability and being allowed to demonstrate it. The distinction matters because an interview is a performance with rules, a time limit, and an audience. Companies need evidence. Candidates need room to produce it. Much of Bhende’s career has unfolded at the meeting point between those two needs, where apparently small choices about process can affect who gets through the door.

Before the interview, a magazine

At Cambridge, Bhende helped start Topic Magazine in 2001. It was a quarterly nonfiction publication whose contributors included musicians, athletes, prison inmates, and competitive eaters. Few corporate biographies make room for competitive eaters. This one should. The magazine reached more than 20,000 readers and attracted attention from The New York Times. Before Karat asked people to show their skills, Topic offered people a place to tell their stories.

Bhende sold issues door to door and gathered information about customers. Publishing gave him a practical education in building a business: someone has to make the thing, someone has to want it, and the distance between those two people must be crossed. The future founder’s first product was something you could hold. His later company would sell something less tangible, a better-organized hour of another person’s attention.

His career also included quantitative analysis at Bates White and management consulting at Oliver Wyman. At Microsoft, he worked in Beijing managing the company’s Corporate China Fund, then led strategic partnerships and global business development for Xbox. Economics, publishing, consulting, investing corporate capital, entertainment: the route refuses to look like a tidy apprenticeship for a recruiting startup. The connecting work was learning how organizations make decisions about people, products, and opportunity.

Xbox had engineers. Who had time to interview?

At Xbox, Bhende encountered a problem hiding inside a perfectly ordinary task. His team needed to hire engineers. Engineers were being asked to interview other engineers. He began counting how much time went into those conversations instead of writing code. The people doing the interviewing had not been formally trained to do it well. An expensive activity was also an inconsistent one, and familiarity had made both defects easy to overlook.

Technical expertise and interviewing expertise have an awkward relationship. Knowing how to build software does not automatically teach someone how to choose questions, listen for reasoning, or compare answers fairly. A good interview asks for evidence that matters to the job. That requires design and practice. Leaving the whole exercise to a busy employee between meetings makes the calendar convenient for the organization while putting an extraordinary amount of weight on an improvised conversation.

Bhende and Jeffrey Spector, formerly chief of staff to Melinda Gates, spent six months traveling around America studying how companies hired software engineers. They heard the same complaints about inefficiency and inconsistency. In 2014 they founded Karat. By October 2018, the company had conducted more than 20,000 interviews for customers including Intuit, MuleSoft, Jet, and Box. “We basically professionalized interviewing,” Bhende said. The phrase is plain enough to conceal how much organizational habit it asks customers to change.

Their model gave the interview a dedicated workforce. Vetted engineers conducted structured conversations by video, working with questions and scoring rubrics. Customers received assessments while their own teams recovered time for other work. The question became whether to build that capacity internally or buy it from a specialist. Bhende had found a business in an activity that companies already performed, often reluctantly, and usually regarded as nobody’s main job.

A second attempt, with fresh questions

The redo is one of Karat’s most revealing choices. In its documented process, candidates can request another interview within 24 hours of the first. They select a new time, meet a different interviewer, and receive new questions. Both recommendations go to the client; Karat recommends reviewing the higher score. The client still decides whom to hire. A second attempt expands the evidence available for that decision.

This is a more demanding proposition than simply being nice to applicants. An additional interview consumes time and costs money. It has to help the employer learn something useful. In May 2022, Karat reported that nearly 1,000 candidates had been hired after a redo and that redos had increased clients’ overall hiring yield by 17% over the preceding year. Those are company-reported results from a particular period, rather than a guarantee for every employer.

A 2022 Wayfair case study put a face on that mechanism through a count: 108 hires from candidates who completed redo interviews. Atlassian later adopted the option alongside round-the-clock scheduling. The details matter. An evening appointment, a weekend slot, or a second conversation changes the demands placed on someone trying to get a job while managing the rest of life. Fairness becomes something a calendar and a workflow can actually accommodate.

Bhende and Spector have also written about the “interview gap,” the distance between the interviewing capacity organizations have and the capacity they need. Their argument links access, consistency, and engineering productivity. A company with too little capacity may restrict who gets evaluated before it has seen what applicants can do. Broadening access to a well-designed interview gives employers more people to consider. It gives applicants a chance to be judged on demonstrated work.

Practice gets a place on the calendar

That emphasis on access became more explicit in Brilliant Black Minds, which Karat launched in 2021. The program provides free mock technical interviews and feedback to Black software engineers, with mentorship, study sessions, and networking opportunities. Its practical premise is easy to grasp: familiarity with the interview process is distributed unevenly. Practice can be arranged. A professional network can be widened. Feedback can arrive before a job offer hangs in the balance.

In April 2022, Serena Williams invested in Karat to support the program and became its “Champion of Brilliance.” The announced ambition was to help double the number of Black software engineers in the United States. That was a goal, not an accomplished result. Williams brought a connection to the discipline of practice that made sense for a program built around repeated opportunities to demonstrate ability.

Bhende’s interest in opportunity also reaches back to his university ties. He served on the Schreyer Honors College Alumni Society Board, helping connect the college with industry and expand students’ career options. In a 2021 conversation with a student founder, he discussed courage, confidence, and priorities. These are modestly concrete forms of involvement: introductions, advice, another chance to rehearse. They sit comfortably beside the much larger commercial operation he leads.

“I co-founded Karat to unlock opportunity.”

Mohit Bhende

When the job changes, the interview must follow

Karat raised $110 million in October 2021 at a $1.1 billion valuation. Tiger Global led the round. Its customers included companies outside the conventional technology sector, such as Ford, American Express, and Bank of America. Hiring software engineers had become a problem shared by businesses with very different products. The valuation marked investor confidence in Karat’s approach at that moment; it was a company valuation, rather than a measure of Bhende’s personal wealth.

The next chapter brought a change in what engineers themselves do. Karat acquired Byteboard in January 2025, adding project-based assessment technology and tools for evaluating work in an AI-assisted environment. Bhende described the combination as a way to measure the skills needed in that workplace. His original concern about an interview poorly matched to its task had acquired a new setting: software development with AI in the room.

In December 2025, Karat launched NextGen Interviews. Candidates tackle multi-file projects with an integrated AI assistant while an expert interviewer probes their reasoning and decisions. The company said its dataset then drew on more than 600,000 interviews. The design allows AI use within the provided environment and examines how candidates work with it. An answer produced with assistance still leaves plenty to discuss: why it works, what it overlooks, and how the engineer checks it.

By April 2026, Karat was publishing lessons from operating the new system, including work on content and calibration. The continuing task is a familiar one for Bhende’s company: keep the assessment connected to the job. If tools change the work, questions and scoring need attention too. The human interviewer remains part of the product, observing a person’s decisions as the tools around that person become more capable.

The Honda has left the building

Bhende’s account of Karat’s early years includes a 2008 Honda CR-V. He and Spector drove it around Seattle, talked through ideas, and accumulated parking tickets. They received verbal confirmation of their Series A term sheet while sitting in it. He kept the car after the funding arrived. By the time he sold it, it had carried him through nearly 75,000 miles of city driving.

In reflecting on the sale, he reconsidered what those parking tickets represented. Urgency had once looked like proof of hustle. He now emphasized the steady control required to continue day after day. It is an unusually useful founder anecdote because its main character is a dependable used car. The milestones happened; the vehicle kept doing its ordinary work. Bhende’s larger project also depends on ordinary work done carefully: scheduling, questioning, listening, giving someone room to try again.