Finding talent is a key job
I joined Veriff as its first ML engineer, and almost immediately we realised we needed to scale up the AI effort very fast. I hired ten people in four months, and ended up leading the team myself four or five months in.
One of those ten was fresh out of university with a master's in physics and almost no professional experience. What I saw was three things: very strong personal motivation and drive, a high baseline capability and intelligence, and a broad technical background in mathematics and programming. I hired him more or less purely on potential, and I was extremely right: after I left he was the one that took over running the entire AI side of the company.
I thought of this recently when a friend of mine, an early-stage founder who spent seven or eight years at a large, prestigious tech company, told me he's having trouble hiring. He never had issues hiring world-class people before. But now it's just really hard to find people.
This starts mattering the moment you can't just hire your friends anymore. When you hire your first, second, and third employee, and then obviously as you scale the company further, finding talent is not a side thing you delegate. It is a key part of the job.
Mediocre people tend to hire people who are worse than them, so B players hire C and D players, and great people looking at your company from the outside will be deterred by poor talent. They will see it in the interview process, and if they don't, they will see it once they join, and then they leave.
Great people want to join a company where the talent is already great. That's its own reason to join, and a very strong one. I think Pactum has done quite well at that, and it's one thing I'm proud of.
Some companies have the luxury of good candidate flow, where great press, a strong tech culture, or very high pay bring in plenty of good candidates without much extra work. In 2026, Anthropic is probably the best example. Tom Blomfield, who co-founded Monzo and was a group partner at Y Combinator, took a leave of absence in July to join their compute team as a member of technical staff. There aren't many companies like that.
For the rest of us, we are competing in a market, and on average you get the market average. That's not that great, and especially in your early team you need extremely strong people. So the question is how you beat the market.
You could pay a lot. But cash is limited, and high pay attracts mercenaries instead of missionaries.
The approach I'm a big believer in is becoming very good at identifying talent. It's not something you can acquire in a matter of two weeks; it has to be honed over time. You need to get good at seeing which people are amazing, or have the potential to be amazing, even when that is hard to read from external signals.
Everyone knows a Stanford-graduated Stripe engineer is probably a strong hire. That person can go anywhere they want. They have their choice of company, they can get extremely highly paid, they can work from any location. Whereas if you're looking at someone with a patchy resume, short stints, no great university, no strong names on the CV, weird titles, gaps, it takes a special skill to:
- go into the large pool of people who don't seem great,
- find the ones who are great but throw off no signal a recruiter could search for on LinkedIn,
- verify them off a short process or a few conversations.
A lot of my best hires are of that kind. Either close to zero experience, or a background that comes off as tangential: academia, the public sector, someone who has never worked as an engineer before.
So how do you get good at this? I don't think there's a way around lots and lots of practice. Over my first year at Veriff I did about two hundred interviews, and there were weeks with twenty hours of interviews in them. You will make mistakes, and you need to make some: it's having lots of data points and acting on them that trains your ability to pick up the signals, whether in person or off a CV. By now I've done over 500 interviews, and been significantly involved in the hiring of about a hundred people who ended up joining.
The most generalizable move I know is to put ten names on paper, people you have worked with who are amazing. Twenty is better, because you need enough of them to generalize from. Then work out which signals were common to them before they were great: how they speak, how they work, where they came from, what their hobbies were. You are pattern matching off people you already respect and personally know, and you can also just ask them. Have lunch, and find out where they came from.
As I've gone through my reps, two resources have been helpful. Talent by Tyler Cowen and Daniel Gross is the more conceptual one. Who by Geoff Smart and Randy Street is the tactical one, on how to run a good process. Being structured is important, although process isn't everything: you still have to get people in the door.
The importance of this skill won't go away. The bigger a company gets, the better you need to be at continuing to hire great talent. Maybe at a thousand people you're not interviewing everyone anymore, but the key people you definitely are, and the skill generalizes to that as well, so it keeps feeding you throughout the life of your company.