George Favvas built for a decade. Then he joined the next class.

George Favvas had already spent a decade running Circle Medical when he returned to Y Combinator. He had a new company, Clara, and colleagues to bring into the room. He also had a perfectly respectable excuse to stay away: experience. In Silicon Valley, experience is sometimes worn like a backstage pass. Favvas chose another turn in the audience, another round of office hours, another opportunity to have somebody ask an inconvenient question.

His explanation was disarmingly practical. Advice could be read online. A conversation about his company's particular problem, in its particular week, could not. And the community would belong to his cofounders, Caitlin Swift and Zeeshan Ahmed, too. Starting again had value even for someone who knew the route. A familiar classroom can still contain an unfamiliar answer.

The decision makes more sense when the preceding years come into view. Favvas's working life has included software development, product management, company formation, investment, teaching and mentoring. Long before Circle, he was helping people compare prices and pool their purchasing power. The titles changed. Questions about customers, distribution and how a business survives kept following him.

Montreal, before the California chapter

Born and raised in Montreal, Favvas studied communications at Vanier College from 1993 to 1996. His studies included design and media, computer science and mathematics; he also participated in the student newspaper and student council. It is a rather agreeable combination for a future product builder: something to make, something to calculate, and somebody who will insist on discussing it.

His early career included TotalNet and the business later known as Radialpoint. By 2006, he was founder and CEO of SmartHippo, which approached financial-product shopping through shared information. Consumers could contribute the rates they had received and compare experiences. The premise gave other people's hard-won knowledge a useful place to go.

In October 2008, Favvas demonstrated SmartHippo at Finovate. The marketplace combined contributions from users, information collected by software and rates supplied directly by lenders. A forthcoming rate engine covered more than 200 lenders. The presentation was about a specific purchasing problem, with an actual mechanism for addressing it. SmartHippo was subsequently acquired by RateZip.

Those details put some texture into the phrase serial entrepreneur. Before the later chief executive title, there was a business built around making a complicated comparison easier. Favvas was already working at the intersection of information, trust and a customer's decision to buy.

In 2015, he described a small California moment that caught the contrast between generations. A Google self-driving car was beside them on the highway, and his son noticed it. To the child, Favvas said, it was normal. For the adult who had moved from Montreal four years earlier, it was still worth remarking upon. The future can be a startling place to relocate; children have a talent for treating it as the scenery.

The move came with a buying group

In 2011, Favvas moved from Canada to San Francisco. Rewardli, which he cofounded with Jean-Sebastien Boulanger, joined the first 500 Startups accelerator class. Real Ventures backed the venture alongside 500. Montreal had supplied both founders and an investor; California supplied the next setting.

Rewardli let small businesses pool purchasing power. The company sought discounts on things they already needed: advertising, office purchases, travel and other business expenses. Members could request deals and vote on requests, giving the company a way to see which negotiations might be worth pursuing. The social network had acquired a shopping list.

At the 2011 demo day, Favvas offered a modest revenue report: “We've already made $100 in revenue using the product ourselves.” The line, recorded alongside laughter, is a useful souvenir from a startup's small beginnings. Before anyone can speak convincingly about scale, somebody has to persuade the first few dollars to arrive.

By October 2013, Rewardli offered a dozen deal categories and more than 500 vendors. Favvas said that some companies reported saving $20,000 to $30,000 a year. One customer, Tinfoil Security cofounder Ainsley Braun, described using it to check purchases and try services. There was a recognisable business underneath the pitch: smaller buyers trying to make a limited budget go further.

The business became PerkHub, providing software for perks and group-buying programmes. Its customer list included American Airlines, TriNet and Google. Favvas's work had moved from organising smaller buyers toward serving organisations that could offer the programmes themselves. The product had evolved; purchasing power remained the subject.

Saturday classes, practical questions

Favvas also taught product management at General Assembly. His course ran for ten Saturdays and covered user experience, testing an initial product and customer acquisition. These are the topics that make a product manager's calendar less glamorous than the title might suggest. A promising idea still has to meet somebody willing to use it.

Teaching sat alongside mentoring at FounderFuel, the Montreal accelerator. He advised founders on product strategy, growth and fundraising. On Clarity, his areas of expertise extended to distribution, affiliate programmes and the practicalities of moving to the United States as a founder. He had made that move himself. Experience could be turned into something another person could use. His public profile made room for questions about customer acquisition as well as fundraising. It is a useful pairing: finding money for a company and finding customers for its product occupy different lines in the same survival plan.

His advice about competition was similarly grounded. A copied interface, he argued, does not give a rival the reasoning or customer knowledge behind the original product. He urged founders to concentrate on the experience they delivered and on finding customers. It is counsel with little theatrical flourish, which may explain its utility. Imaginary competitors have an inexhaustible capacity to attend meetings.

A 2013 public comment to the SEC showed another concern from the same period. Favvas welcomed efforts to ease startup fundraising but objected to burdensome registration requirements. He was paying attention to the mechanics around the company as well as the product inside it.

Ten minutes, then across the street

In 2015, Favvas cofounded Circle Medical with Boulanger and Brent LaRue. Boulanger was already a collaborator from PerkHub. LaRue later recalled that he and Favvas first met through a General Assembly class. The teacher's timetable had produced a connection with consequences well beyond the course.

Their Y Combinator interview required a second appointment. After the first group deliberated, the founders were sent across the street to another group. Favvas remembered Sam Altman sitting on his side of the table in a Reddit T-shirt, leaning forward and attending closely to their answers.

The business was still pre-revenue. Favvas did not remember their pitch as particularly polished. He remembered rapid iteration and close attention to users, and he remembered the intensity of the questioning. The team entered YC's summer 2015 batch. Years later, he could still describe the seating arrangement. Some business encounters leave a more durable impression than their slide decks.

“Community. This is the gold.”

George Favvas, on returning to Y Combinator

The arithmetic of staying

Favvas remained CEO for a decade. When he announced his departure from the role in October 2025, he described revenue run rate rising from $2 million to $114 million over five years, alongside profitability and $12 million in primary capital. Run rate is an annualised pace, so those figures describe the scale reached at particular points, rather than a claim about every intervening year.

He also described a team of more than 200 people, with colleagues in San Francisco, Montreal and elsewhere. The geography had come around again. The city he left remained part of the organisation he built. His departure message included another measure of its afterlife: four alumni had started companies, including two backed by venture investors.

For the colleagues who stayed, he stressed values, long-term decisions and doing the right thing when nobody was watching. He offered to remain available. The farewell had room for financial results, but it also addressed the people who would inherit the ordinary decisions after the departing CEO's announcement had slipped down the feed.

Back in the room, with Clara

By May 2026, Favvas was launching Clara with Swift and Ahmed. The company announced a $12 million pre-seed round led by Y Combinator and A.Capital. Liquid 2 Ventures and a group of individual investors also participated. Swift and Ahmed had worked at Circle, bringing another set of existing relationships into a new venture.

Clara joined YC's spring 2026 batch. Favvas's comparison with 2015 was specific: roughly 100 companies then, about 200 this time; a demo day once held at the Computer History Museum, now spread across three YC buildings with streets closed to traffic. He likened the newer event to a block party. Even the business of beginning had acquired a larger guest list.

What he valued in the programme remained familiar: questions in office hours, a community of founders and the concentrated activity around demo day. Clara had raised money before the batch, yet he still considered attendance worthwhile. His argument left room for an experienced founder to need other people, and for his cofounders to build relationships of their own.

Favvas's career has crossed enough company names to make a lengthy introduction possible. The more interesting detail is his willingness to return to a setting where introductions are only the beginning. After the move, the demos, the Saturdays spent teaching and the decade in charge, there was another company to build. He took his place in the next class.