The playbook for building high talent density teams | Adam Ward, Head of Talent at Cursor

My guest is Adam Ward, now leading talent at Cursor after co-founding Growth by Design, the recruiting firm Cursor acquired. He previously scaled recruiting at Pinterest and Facebook. I wanted to get tactical about identifying remarkable people, persuading them to join, and not building an average team by accident.

This hiring market is an eleven. It rhymes with the mobile scramble, when everybody needed mobile engineers before the category really existed, but this cycle is compressed into days and weeks. It’s a tale of two cities: enormous offers for new PhDs beside layoffs at terrific companies. Talent, labor, and how work gets done are resettling in real time. Forward-deployed engineers are especially scarce: deeply technical people who can work with customers and sales, make a complicated product legible, and help get it into a real workforce. Narrow specialists are softer. Engineering, product, and design are squeezing together, so taste, judgment, curiosity, product sense, and systems thinking matter more. AI can generate plenty of slop; I’m optimistic it brings us back to craft.

You call the conventional approach the funnel of doom. Companies gather a huge pile of applicants, referrals, and sourced people, screen them down, then hire whoever remains. But if we want the best people, why would the output of a weeding-out exercise be the answer? How do we hire more like executive search, starting with the handful who could genuinely be exceptional?

The first move is scope. “I’ll know it when I see it” is not a strategy. Before speaking with candidates, stack-rank the abilities and experiences that matter, why the role exists, what success looks like, and why it will have impact. Don’t begin with resume logos. Then build a market map. For a senior search, maybe fifty people can do the job; go find those fifty. Ask people with real taste specific questions: who collaborates unusually well with design, or turns an abstract framework into a product better than anyone? Those prompts map to your scope, and names triangulate. Find companies that value the same traits, not just the usual prestige list.

Tools help, but the highest-value work is personal: sit with someone repeatedly, work through their network, follow leads, and ask about the person they forgot last time. Founders can’t say talent is number one and hand it away. Recruiters are confidence engines, not decision engines; managers need to own hiring. At Cursor, someone drops an impressive project or person they heard about into a hiring-ideas channel, and the company rallies: who knows them, what might interest them, how do we connect? That beats asking the room for a random profile.

Relentless pursuit feels slow at day zero because seeds can take months or years. But “not now” doesn’t mean disappear until late fall. Get coffee, visit the office, meet someone fascinated by their work. Learn what may engage a person before you write. Once somebody ignores your first attempt, the next ones get easier to ignore. Caring is free: people remember that this company truly wanted me, and I could belong there. We’re trying to hire one person, not optimize ratios. Make it bespoke—match them with a teammate who shares their interests, eat together, and bring them in when it feels right.

Work samples are among the strongest predictors we have. For some roles, we use extended, side-by-side project experiences so candidates gather data about us too. An onsite should not be pure extraction. Curate who they work with, who greets them, even who sits beside them at lunch. The recruiter quarterbacks the experience, but the hiring manager drives the substance. We tried removing work trials because they’re expensive and a big ask; our confidence dropped. Debrief that day or the next. The manager gets input, then makes the call, because they own the outcome. When someone is ready to move, move fast.

Closing starts long before an offer. Understand why someone might leave, what they value, and what makes the role meaningful; make the whole process speak to that. We’re trying to win hearts and minds, not buy our way out of a bad fit. Compensation is wild because time is scarce, but a huge number can become someone’s new normal while they remain unexcited by the work or people. Closing is a team sport. Personal details matter: for a classically trained violinist, we found a distinctive stringed instrument for the offer dinner. It said we paid attention. That thoughtfulness works when you’ve deliberately built relationships with a small group.

Early companies often make a silver-bullet first talent hire: fill ten urgent roles and build the whole recruiting system. Those are competing jobs; decouple them, even if temporary help handles one side. I want recruiters curious about the business, aware of how each hire changes its slope, and with a high motor. Talent density is collective, not one mythical ten-x hero. Great managers are puzzle builders. An exceptional individual gets throttled on a weak team; I’d rather build a team that becomes exponentially better together.

After acceptance, don’t just move to the next candidate. Reneges are real, especially with a long gap before joining. Keep building community with dinners, touchpoints, and pre-boarding. My warm take on the talent engineer is that every recruiter should become one: we can build useful systems ourselves now. Parenting is humbling and my greatest joy. The best career moments are often when you’re over your skis—unprepared, uncomfortable, figuring it out through ups and downs. Raising kids has that shape too. And honestly, I’d love people to try Cursor. Things are changing quickly; builders need the best tools for the task in front of them. I’m always down to be helpful.