AI in Hiring: How Kevin Turned 200 Mock Interviews into a Human-Centered Interview Platform

AI in hiring

AI in hiring has a reputation problem.

Candidates feel like they are shouting into a black box of automated rejections. Recruiters are overwhelmed by thousands of résumés and rushed screenings. Companies waste time and still miss out on incredible people.

In this episode of AI Inside San Francisco, I sat down with Kevin Hsieh, CEO of Mirwork and a product leader turned founder, who has lived every side of that story. He sold gaming PCs as a teenager, built medical devices, joined big tech after 200-plus mock interviews, and now is building an AI interview platform designed to make hiring more human, not less.

“I don’t really want to replace either side,” Kevin told me. “At the end of the day, it’s a human working with a human, not a robot working with a robot.”

This is how he went from a one-way ticket to the US to rethinking AI in hiring.

https://www.youtube.com/watch?v=b6fK2e3_W4Q

From One-Way Ticket to Finding Home in Tech

Kevin’s journey into AI in hiring starts long before AI existed.

His parents sent him from Taiwan to the United States in middle school, with a one-way ticket.

“When I first moved to the United States, I didn’t know I was going to stay here,” he said. “I didn’t even clean out my stuff because I thought I was going back to Taiwan.”

He landed in Wisconsin, barely speaking English, sitting in classrooms where he understood almost nothing.

“Going to class was not even fun for me, because I had no idea what everybody was talking about.”

Two forces changed that: a math teacher and a house full of computers.

“There was a math teacher who really inspired me. Her name is Mrs. Aschta,” he recalled. “She really encouraged me to pursue what I like. That’s when I felt like, ‘Oh, maybe the US is not that bad.’”

At home, his dad taught computer science and kept multiple machines around.

“My home would have five or six computers growing up,” Kevin laughed. “So I definitely blame my dad for making me into a gamer.”

That support, plus all that hardware, became the foundation of his first side business.

The Side Hustle: 300 Custom Gaming PCs and a Lesson in Value

By high school, Kevin was building and selling custom gaming PCs and accidentally learning the basics of product and customer empathy that would later shape his view of AI in hiring.

“I was fascinated about buying the fastest RAM and the biggest graphics card on the market,” he said. “I’m pretty sure I built almost 300 custom computers.”

Interestingly, he was not cold-selling. He was hosting.

“A lot of times people bought the computer not because I tried to sell to them,” he explained. “I liked hosting. I’d cook, have a small party at home, and people would ask, ‘Why do you have six computers in your living room?’”

From there he would help them figure out what they actually needed.

“I learned early that when you want to sell something, you actually want to create value for people,” he said. “You empathize with each user and understand what they really want.”

That lens — start with empathy, then technology — is exactly what he later applies to AI in hiring.

From Saving Stroke Patients to Building Products

Originally, Kevin wanted to be a doctor.

“As a kid, I always wanted to help people,” he said. “I even tried to bandage myself and look very professional.”

He studied biomedical engineering at UC Irvine and later at National Taiwan University, working on medical devices such as ultrasound systems for stroke patients.

“We were building devices where you only have six golden hours to help a stroke patient,” he said. “It’s super impactful, but also very hard emotionally.”

Working in hospitals made him rethink where he could have the most impact.

“I started wondering, can I help people before they get to the hospital?” he said. “Can we do more preventative work and create value earlier?”

That drove him toward product management and tech, including smart home systems and AI-related healthcare tools. One of the most impactful: an autism screening solution.

“We were the first software company that got FDA approval for autism screening and diagnosis,” he explained. “We cut the wait time for parents from one year to one week.”

The pattern kept repeating: use technology to augment human decisions, not replace them. That same principle now shapes his approach to AI in hiring.

200 Mock Interviews and the Brutal Reality of Big Tech Hiring

Even with that background, Kevin felt his resume lacked one key ingredient: big tech experience.

So he set his sights on Meta.

“I’d never worked in FAANG before,” he said. “I felt that exposure was the missing piece.”

A Meta recruiter reached out to him first because of his AI background, but that did not make the process easy.

“The more I studied, the more I felt underprepared,” Kevin admitted. “So I decided: when I walk into that interview, I don’t want to regret not preparing enough.”

What followed was intense:

  • 200-plus mock interviews

  • 400–500 hours of practice

  • Dozens of partners and feedback sessions

“I spent at least 4 or 500 hours trying to find people, schedule mock interviews, reschedule, and build awareness of my strengths and gaps,” he said. “I wanted to be the best version of myself.”

He got the role. But going through that process exposed a deeper systemic issue with hiring that AI alone has not fixed:

  • Candidates spending nine months or more searching for a job

  • Thousands of applications sent with almost no human interaction

  • Interview processes that feel random and opaque

“Some people inspire me,” Kevin said, “but at the same time I feel very sad because how can such a wonderful individual take nine months to find a job and go through a thousand applications?”

That pain led him to rethink how AI in hiring should actually work.

The Real Problem: Compatibility, Not Just Qualifications

Kevin challenges the standard narrative that hiring is only about skill and experience.

“I want to emphasize compatibility more than qualification,” he told me. “Qualification is often subjective. Compatibility is about: do we actually work well together?”

He compares hiring to dating.

“It’s like dating — you can list all the traits you want,” he said. “But when people meet, if they like each other, a lot of the other details become less important.”

He also calls out how broken the front end of hiring feels for candidates, especially in the era of AI in hiring:

  • They fear that an automated system silently rejects them.

  • They rarely get to speak in their own voice.

  • They almost never get clear feedback.

“One big myth is this idea that there’s an AI overlord that just slashes candidates as pass or fail,” he said. “That’s not how good hiring should work.”

His answer is not less AI, but better AI.

A Different Approach to AI in Hiring: An Interviewer That Listens

Kevin’s startup is an AI interview platform built around a simple idea:

👉 Use AI in hiring to create more opportunities for humans to connect, not fewer.

There are two main flows.

1. Hiring Experts Design the AI Interviewer

The platform lets recruiters and hiring managers configure the AI to match their culture and process:

  • Question types and depth

  • Evaluation criteria such as communication or analytical skills

  • Tone and style of interaction

  • Even the voice of the AI, including the option to clone their own

“The hiring experts are the more experienced people,” Kevin said. “They should instruct what the AI does on their behalf.”

The result is not a generic robot, but a custom AI interviewer anchored in the team’s actual expectations.

2. Candidates Interview on Their Own Time

Once the setup is done, candidates can join an AI-led interview via a link, often embedded in the job application.

“We tell them this is an AI interview and it’s there to help you share more about yourself beyond the resume,” he said.

Instead of a form, they get a conversational experience closer to talking with a human.

“If you look at a résumé, it’s a one-way interaction,” Kevin said. “In a 30-minute conversation you get thousands of words, all iterative. That’s very different.”

For candidates, the benefits are:

  • A chance to actually be heard

  • More context than a bullet-point CV

  • Faster outcomes, even if the answer is no

“A timely rejection is better than a black box,” Kevin emphasized. “Candidates want to feel like their voice was heard, even if the answer is no.”

Inside the Analytics: Signal, Not Verdicts

On the back end, the system uses AI to analyze interviews and support better decisions in hiring:

  • Word choice and tone

  • Communication clarity

  • Logical structure and problem-solving

  • Criteria defined by the hiring expert

However, these scores are not final judgments.

“We present more information to the hiring team,” Kevin said. “They can override any metric. We’re not there to say ‘the AI is right’ — we’re there to give them signal.”

Human reviewers can adjust scores, leave notes and use the interview data as one input among many. AI in hiring becomes an assistant, not a judge.

Bias, Ethics and Human Oversight

Kevin is blunt about bias.

“I believe everything has a bias — humans and models,” he said. “The question is: how do we make it more consistent and accurate, and keep humans in the loop?”

His strategy:

  • Standardize what questions are asked and how they are evaluated

  • Document the process so decisions can be audited

  • Keep humans in control of final decisions and overrides

“We’re not here to judge people or tell companies what culture they should have,” he said. “We want to enable both sides to showcase their true value and make decisions faster.”

In other words, AI in hiring should support fairness, not pretend to be perfectly objective.

The Biggest Myth About AI in Hiring

When I asked Kevin about the biggest myth in AI hiring, he did not hesitate.

“The biggest myth is that there’s this AI algorithm that just slashes candidates pass or fail,” he said. “Especially after talking to so many teams, that’s not how it really works.”

Most hiring teams still manually review résumés and interviews. The real issue is not some evil AI gatekeeper, but limited attention and messy workflows.

“The problem is more about whether that 10 seconds of attention is seeing the right signal,” he explained. “We want to help them see more of the right things in less time.”

Again, AI in hiring should be a lens, not a wall.

Lessons for Candidates and Companies

There are a few clear takeaways from Kevin’s journey.

For Candidates

  • Preparation matters. A lot.

  • Mock interviews and feedback loops are game-changing.

  • Focus on clarity, not just complexity.

“I wanted to walk into the interview and know I did my best,” Kevin said. “I didn’t want to regret not preparing.”

For Companies and Hiring Teams

  • Don’t hide behind tools. Use AI in hiring to hear more voices, not fewer.

  • Design your AI interviewers to reflect your real culture.

  • Use analytics as input, not as the final word.

“We’re trying to use technology to help humans meet each other faster,” Kevin said. “Not to replace either side.”

Why Human-Centered AI in Hiring Is the Future

When you strip this story down, Kevin is doing what he has always done:

  • As a teenager: help people get the right computer, not the most expensive one.

  • As an engineer: build tools that help doctors make better decisions.

  • As a founder: create AI in hiring that listens first, then assists.

“We want to make the AI experience more encouraging and more empathetic so we can get more genuine information from both sides,” he said.

If AI in hiring is going to work long term, it needs exactly that balance: technical rigor with human empathy.