The companies winning AI talent are not the ones with the biggest comp packages. They are the ones who can articulate why this specific role at this specific moment matters. Vercel's Amina Moinuddin Darwish runs the playbook that turns recruiting into a CEO-level priority, not a TA function operating in isolation.
Amina Moinuddin Darwish, Global Head of Talent at Vercel (the company behind Next.js and a top destination for AI-native engineers), joined 10x Recruiting (the show) host Nolan Church to explain how Vercel is winning hiring in the most competitive talent market in tech. Her three structural moves: evaluate recruiters on the long-term impact of the hires they make, gate every offer through CEO approval to preserve talent density, and apply Vercel's "iterate to greatness" mantra to recruiting itself.
What follows is the operating model Amina runs at Vercel: how the war for AI talent actually gets won, why long-term hire impact (not time-to-fill) is the only metric that matters for recruiters, how a CEO-approval gate works without becoming a bottleneck, and the iteration loop that compounds recruiting quality every quarter. The throughline: recruiting at the best AI companies is a craft, not a funnel.
Winning the war for AI talent
Every AI company is recruiting from the same shortlist. The candidates who can build frontier inference infrastructure or fine-tune large models number in the low thousands globally. Comp packages have converged. The signing bonus war has ceiling effects.
What separates the companies that close those candidates from the ones that lose them is narrative discipline. The candidate is choosing between three offers that all pay roughly the same. The deciding factor is which company can articulate, in a way that lands with that specific person, why the work matters now. Vercel's recruiting motion is built around producing that articulation per role, per candidate, per stage of the conversation.
The candidates we want are choosing between companies they could all be successful at. Our job is to make sure they see why being successful here, in this specific moment, is the most interesting bet they can make.”
The recruiting team's job becomes translating Vercel's product trajectory into language the candidate already cares about. The model is closer to enterprise sales than to traditional sourcing. The recruiter spends the first call understanding what the candidate is actually optimizing for, then reframes the opportunity around it.
Recruiters evaluated on long-term hire impact
Most recruiting teams measure recruiters on time-to-fill, offer-acceptance rate, and pipeline volume. Vercel measures recruiters on whether the hires they made are still moving the company forward 12 to 18 months in. The metric is uncomfortable because it lags. The signal it produces is the only one that actually matters.
The mechanism is simple: every quarter, the recruiting team reviews the hires from prior quarters against the outcomes those hires were brought in to deliver. Recruiters who consistently bring in high-impact hires get more responsibility. Recruiters whose hires churn or underperform get retrained or rotated. The accountability loop is honest because the data takes long enough to surface that nobody can game it.
The side effect is that recruiters become much more invested in qualification. A close-rate optimiser will push a borderline candidate over the line. A long-term-impact optimiser will walk away from the same candidate because the downside risk is on their scorecard 18 months from now.
The CEO approval gate
Every hire at Vercel goes through the CEO. Not for rubber-stamping. The CEO actually reads the offer summary, sometimes meets the candidate, and can veto. The gate sounds like it would slow hiring to a crawl. In practice it does the opposite.
The gate produces three downstream effects. First, hiring managers stop bringing borderline candidates to offer because they don't want to defend a weak hire to the CEO. Second, candidates take the process more seriously because they know the CEO is paying attention to their candidacy. Third, the company maintains a coherent picture of who is joining and why, which keeps talent density from drifting as the team scales.
Iterate to greatness, applied to recruiting
Vercel's company mantra is "iterate to greatness." It started as a product principle and became an operating principle. Recruiting runs on the same logic. Every search teaches the team something about what works and what doesn't. The team captures the lesson, adjusts the process, and applies the refinement to the next search.
The mechanism is a weekly recruiting retro. What worked this week. What didn't. What we're changing for next week. The retro produces small adjustments compounding into a recruiting org that operates materially better quarter over quarter. The opposite pattern (annual playbook updates, pre-baked frameworks) decays under contact with the actual market.
Recruiting is a craft. You don't get great at it by following a playbook. You get great at it by running searches, paying attention to what worked and what didn't, and adjusting the next one. The teams that compound are the ones that take the iteration loop seriously.”
Talent density as competitive moat
The phrase "talent density" gets used loosely. Vercel uses it precisely. Every new hire raises or lowers the average bar of the team they join. The 50th hire becomes the bar that the 51st has to clear. Maintained over hundreds of hires, the bar compounds upward. Allowed to slip, it compounds the other direction.
The CEO approval gate exists to enforce this mechanic. The long-term-impact recruiter evaluation exists to enforce this mechanic. The "iterate to greatness" loop exists to enforce this mechanic. Each piece looks like a separate operational choice. They're actually the same choice expressed in different surfaces of the org.
The competitive moat is not the comp package or the office or the equity. It's the answer to the question every senior candidate asks: "who would I be working with?" When the answer is consistently strong, the candidate signs. When it's mixed, the candidate keeps interviewing.
Where AI gives recruiting teams use
Vercel's recruiting motion is intensive. Per-candidate narrative work, weekly retros, CEO-gated offers, long-term hire reviews. The depth is what produces the outcomes. AI is what makes the depth possible without doubling the headcount of the recruiting team.
The use points: Metaview's Notetaker captures every candidate conversation so the recruiter can be present in the room instead of taking notes. AI surfaces patterns across the conversation history that no individual recruiter would spot. Application Review handles the structured intake on application volume so the team's hours go to the candidates who actually need senior recruiter attention.
Across 505 recruiting and hiring leaders surveyed in Metaview's 2026 AI & Hiring Alignment Report, the teams running AI as core to their hiring process pull away from the teams that don't. The data:
That last number is the one that connects to Amina's playbook. Hire-to-role fit is exactly what the long-term-impact metric measures. AI captures the conversation data that makes the 18-month retrospective possible.
The operating shift
Amina's three structural moves translate cleanly for any recruiting team trying to win in a tight talent market:
One: switch the recruiter scorecard from time-to-fill to long-term hire impact. Review hires at 6, 12, and 18 months against the outcomes they were brought in to deliver. The recruiters who consistently produce winning hires get more responsibility.
Two: install a CEO or named-veto gate on every offer. The gate matters more than who runs it. The signal it sends to candidates and hiring managers compounds.
Three: run a weekly recruiting retro. What worked, what didn't, what you're changing next week. Compounding small adjustments beat annual playbook rewrites.
The companies treating recruiting as a craft are the ones who keep winning the talent they need to ship the work that matters.
Bring Metaview into your hiring stack.
Live notes, structured scorecards, and ATS sync - set up in under 10 minutes.
Frequently asked questions
How does Vercel win against bigger AI companies for the same candidates?
By out-narrating them. Comp has converged across the top AI companies. The deciding factor is which company can articulate why this specific role at this specific moment is the most interesting bet the candidate can make. Vercel's recruiters spend most of their time translating Vercel's trajectory into language each candidate already cares about.
How are recruiters evaluated at Vercel?
On the long-term impact of the hires they make. Time-to-fill and offer-acceptance rate are tracked but not central. The actual evaluation happens 12 to 18 months after the hire, when the team reviews whether those hires are still moving the company forward. Recruiters whose hires consistently perform get more responsibility.
Doesn't a CEO approval gate on every hire slow recruiting down?
In practice it does the opposite. Hiring managers stop bringing borderline candidates to offer because they don't want to defend a weak hire. Candidates take the process more seriously knowing the CEO is paying attention. The gate keeps talent density coherent as the company scales.
What does "iterate to greatness" mean for recruiting?
A weekly retro: what worked, what didn't, what we're changing for next week. Compounding small refinements produces a recruiting org that operates materially better quarter over quarter. The pattern beats annual playbook updates because the playbook decays under contact with the actual market.
Why does talent density matter so much?
Every hire raises or lowers the average bar of the team they join. Maintained over hundreds of hires, the bar compounds upward and becomes a competitive moat. Senior candidates' deciding question is "who would I be working with?" When the answer is consistently strong, they sign.
