
Chime’s Matt Valentino on Recruiting When Every Resumé Looks Perfect
AI has made a polished application nearly meaningless. Chime’s VP of Talent and Experience explains what should replace it, where recruiting still needs a human, and why a bigger pipeline isn’t the goal.
Matt Valentino started recruiting when resumes arrived by fax. He watched online job boards and then LinkedIn change who companies could find and how quickly. He thinks AI is the biggest shift since then. This time, though, the change is less about finding candidates and more about evaluating them.
When any candidate can use AI to tailor a resume to a job description, the application stops telling you much. That forces talent leaders to reconsider how they assess people, what they measure, and which parts of the process still need a person.
Valentino has more than 20 years of experience to draw on. He built teams at Genentech that developed first-in-class cancer therapies and held leadership roles at Calico Labs, Alphabet’s longevity research company. He helped DoorDash grow from 2,000 to 17,000 employees before its IPO and through the pandemic, then served as Chief People Officer at Shape Therapeutics.
Today he’s Vice President of Talent and Experience at Chime, where he oversees recruiting, talent management, and workforce planning.
In this Q&A, he explains how to tell whether AI is improving hiring decisions or just adding volume, where automated communication should give way to a person, which investments growing companies make too early or too late, and how he’d build a talent acquisition function today with a small team and AI.
Q&A
You’ve compared AI’s impact on recruiting to the arrival of LinkedIn. What’s fundamentally changing about how companies find and evaluate talent, and which assumptions about recruiting do you think will no longer hold?
I started recruiting when people sent resumes by fax. The move to online recruiting, and then to LinkedIn, completely changed what candidates you could find and how fast. AI has been the biggest shift since then, and what it changes most is the evaluation process.
If I were job hunting today, I’d give AI the job description and my resume and have it make me the perfect fit—so a polished application no longer tells you much. Instead, interviews and real-life examples and work give us more clarity. And even those are changing.
For example, our engineers no longer do a traditional coding session during the screening process. Instead, they do an agentic coding exercise so we can see how they actually work with AI tools.
As AI makes it easier to source, screen and engage candidates at scale, how should talent leaders determine whether that additional recruiting activity is improving the quality of their hiring decisions?
More activity isn’t the goal. If AI triples your pipeline and you hire the same people you would have anyway, you’ve just created more work for your team. The tools are remarkable, we could turn on AI-led first-round screens tomorrow.
But the measure of success isn’t how many candidates moved through the funnel. It’s whether the people you hire raise the bar on their teams, thrive in a fast-moving, in-person environment, and are still performing a year or two later.
If those outcomes aren’t improving, the extra activity is noise, and talent leaders should be willing to say so.
You’ve argued that the human touch in offers and rejections becomes more important as recruiting becomes automated. Where should companies draw the line between automated and personal communication, particularly when they’re managing hundreds or thousands of applicants?
Let automation do what it’s good at. Sorting and reviewing thousands of resumes is one of those things, and an agent might do it better than humans do. The line, for me, is investment. Once a candidate has invested real time in you, they deserve a person.
Anyone who’s interviewed with us shouldn’t get a robo-call telling them they didn’t get the job. The offer and negotiation stage is deeply personal, almost the psychologist part of the job, and that’s where a human matters most.
The same goes for getting people into the office, because seeing the energy and pace tells a candidate more than any careers page. The more senior the candidate, the more this matters. Packages are competitive everywhere, so what tips the decision is when they can say they “felt really good about that team.”
You’ve helped build talent functions in organizations at very different stages of growth. What changes about the recruiting model as a company scales, and where do you see growing organizations make investments too early or too late?
I’ve recruited at a large biotech, at an early-stage research company, at DoorDash during its pandemic hypergrowth, and now at Chime as a public company, and each stage needs a different model.
The most common mistake I see is adding management layers too early. Companies hire leaders whose whole job is overseeing other managers, when what they really need is more people doing the work. The strongest teams keep leaders close to the work. We’ve gone the other way, with leaders who pull their weight and engineering line managers who are now tech leads still shipping code.
What companies get clear on too late is who they are. When I was in the biotech field, nobody had to say that we were “mission-driven” because the work spoke for itself, and the culture of being in the lab together carried over to everyone. If you’re advertising unlimited PTO and free lunches, you’ll attract people who leave the minute someone offers a sabbatical.
For example, Chime is intentionally an in-person, high-intensity environment focused on changing the way banking serves Americans, and that clarity brings in the people who actually want to be here. It also allows us to attract drivers and builders who want to make a big impact on our mission.
If you were rebuilding a talent acquisition function today with a relatively small team and access to AI, what would you design differently from the traditional recruiting model, and what would you deliberately preserve?
A small team today can do what used to take a big one, so I’d design for leverage. I’d let AI handle sourcing, sorting, and early screening, test every candidate for AI fluency, and keep the team flat, with leaders who carry a real workload alongside their team.
I’d also rethink the recruiter’s job: the time AI saves on screening should go into what only people can do, which is understanding what a candidate really wants and helping them make the decision.
What I’d deliberately preserve is the human part. That means getting candidates into the office, being honest up front about the pace, and having leaders show up personally for the hires that matter most. Hiring people who want that environment is how you get people doing the best work of their careers.



