Some interviews take time. This one was more than a decade in the making.
I first met Eric Glyman and Karim Atiyeh in college. A few years later, when they launched a service called Paribus to help consumers get money back from purchases where prices dropped, I was an enthusiastic early sign-up.
Fast forward more years than any of us would probably care to admit, and they’re the co-CEOs of Ramp, the $60 billion startup that’s become one of tech’s highest-profile startups, and the most valuable in New York.
For the nearly 100,000 companies that use Ramp, it’s still best known for its corporate cards that can help track and reduce expenses. For everyone else – the highly online on Tech Twitter, or everyday people walking through Union Square on a given day – it’s the company that flexes a seemingly endless reserve of budget and creativity for marketing stunts with the likes of Kevin from ‘The Office,’ Tony Hawk (he “loves ramps”) and even a 70-minute Broadway musical about paying bills.
Is there a method to this madness? How do we reconcile Ramp the stunts company with Ramp the efficiency-obsessed tech darling, where office scoreboards blare how much a customer is saving?
It’s a question I was especially excited to ask Glyman when he came on this week’s podcast, an interview for which, intentionally or not, I’ve been doing my homework now for many years.
“You can have fun,” he says. “I think the biggest battle that a lot of upstarts and startups face is people not giving a shit about what it is that you’re doing, and why. So there’s this question of, ‘How do you be interesting?’”
We cover so much more ground in this episode of The Upstarts Podcast.
Glyman shares why he’d rather run Ramp like Amazon than Google; why Ramp hires for proof of work, not credentials, from a 14-year-old intern to a 28-year-old CTO; and how being the first customer for Anthropic’s Claude Code and Cognition’s Devin led Ramp to build its own AI tools.
Plus, he shares the real story of what happened to that first startup, Paribus; what he and Atiyeh learned at acquirer Capital One; and how it can serve as a warning for Instinct and Muse today.
My favorite highlights for startup builders below:
A MESSAGE FROM OUR SPONSOR
Move from intent to impact
Jira makes it easy to capture tasks from Slack, turn your ideas into structured plans, and hand off work to agents — in Jira or from where you work. Plus, it taps your team's context so you never start from ground zero. Run your work from your favorite AI tools with the Atlassian MCP or CLI.
How Ramp ships product
Ramp has two main routes to shipping a new product: external demand, and internal capability. Both are meant to follow the same basic directive:
“If a company chooses to spend $100 in order to get $200, how can we improve that ratio?”
Ramp moves the most money through bill payments today. That feature came out of sustained customer demand for it, Glyman explains.
Other new products, like Glass and Inspect, Ramp built initially for itself. Inspect, an AI agent tool, is particularly interesting, since Glyman notes that Ramp was privileged to be the first alpha customer of both Claude Code and Cognition’s Devin, and one of the first 10 customers of Cursor (now part of SpaceXAI).
Ramp’s engineers are “really opinionated people,” Glyman says, and one of them wanted a different form factor and built it themselves. That spread across the organization, and then a blog post about how they did it went viral. Customers asked for it, and eventually they shipped it.
What would make a popular internal tool not something Ramp would ship to customers? The biggest determinants, Glyman says:
Will customers encounter this problem?
Is there a reason Ramp should be better?
I was also curious how Ramp juggles a high-performance culture with experimentation.
The key, Glyman says, is encouraging engineers to focus on building a system that drives an outcome, not a rigid tool. That system can then be approached creatively.
“A lot of this is actually trying to create a culture where people care about the outcome, and giving people the flexibility and permission to work on things that they, in their own judgment, think actually might be a better way to do it.”
He contrasts the product strategies at Amazon and Google: Google gives you 20% of your time to experiment, and merges in a product that loses out to another head-to-head. AWS might have nine products overlapping, explicitly competing. “Sometimes it’s a total zoo,” Glyman says.
But Ramp would rather be like Amazon.
“I think if you have like small teams, five, six people, everyone knows who’s making the call, how are you measuring it. If something is off, you can push each other. And one of the best ways you could make that team unproductive is if you just add 30 people and say, ‘Go agree with each other.’
If you look really deep into Ramp’s organizational design, it’s tons of small teams who say, ‘Here, here’s what we want you to do. Here’s the outcome we want you to drive, but it’s up to you. You can make the call on how you wanna go and spend your time.’
Some people have ideas that will surprise you with how useful and brilliant they are.”
How Ramp hires killers
What kind of employees flourish in that kind of environment?
Ramp looks to hire people who have built or created something of their own and shown initiative above all, Glyman tells me.
“I’m always curious when people start their own, like, clubs in high school. I’m always curious when people have built their own products.”
Ramp hired a 14-year-old intern recently, Glyman adds, not for the shock factor, but because the teen had built his own app that they saw lots of people using on Twitter.
“I think we do try to select for people who’ve shown a bias towards action, and, have shown some proof of work. And sometimes we’ll miss people who have great potential, but haven’t shown that yet.
But I think one of the best ways is to find people who have that drive and then just give them resources. Unblock them. Help them achieve that. It’s what I think my job is.”
Along those lines, one of the reasons Ramp elevated Atiyeh to co-CEO after a number of years was because it wanted to promote a 28-year-old star former founder it had acquired, Rahul Sengottuvelu, to be CTO.
“No one really has more than five years of experience with AI, but he was there, and he was building a product. Probably the closest things that are related to it are Sierra or Decagon.”
Why Ramp’s founders stay hands-on
When I ask Glyman how he and Atiyeh stay hungry and motivated, now they’re a heavyweight and not so much a challenger, he doubles down on how “early” Ramp still is.
Most people in the U.S. still haven’t heard of Ramp. Most card volume, bill payment volume, and treasury deposits don’t go through Ramp.
“We haven’t done much of anything, you know? There’s millions and millions of businesses that we haven’t done a single thing for.”
Among the companies that do work with Ramp, it’s improved how much money it saves a business, from 2% in 2020 to 7.5% in 2026. “I think we can do a lot better. And there’s a lot more great products to build,” Glyman adds.
Glyman and Atiyeh also don’t believe in hiring professional managers, or becoming them.
“We still write copy, we still write code. I walk basically the office floor every day, saying hi to people. I try to go and jam, see what we are working on in designs. I visit customers. I think actually just spending time staying humble enough to really understand what is the work, to do it yourself, to be in the shoes of people [is important].”
Austin Hughes, a former Ramp employee who is now the CEO of Unify, which we wrote about last year, told me in prep that Glyman reading and invoking business books during All Hands meetings left a big impression.
In the episode, Glyman mentions learning from books by famous sports coaches Bill Walsh and John Wooden. One thing these leadership works taught him: his job is to make employees feel like the hero in their stories.
“Everyone wants to be able to look back and say, ‘The work I did mattered.’
And the things you remember, I think, are not the ones that were easy, but it was hard, but, you know, ‘I reached the top of the mountain. I’m so glad I got there. I won the race. I pushed myself to a level that I didn’t know I could get there.’”
Stories of leadership and coaching help Glyman feel less overwhelmed by the number of choices and tasks he faces every day.
“I think one of the things that’s a little hokey now, it probably sounded more normal in the 1950s, but John Wooden would use this phrase of: ‘Make every day your masterpiece.’
I think there’s something really beautiful about it, right? There is so much more that people can do, and I think that there’s a level of performance that you didn’t know that you’re capable of.”