The machine arrived at a video rental store outside São José dos Campos, Brazil, with the charisma of office furniture. It was an Intel 286, bought for the business run by Marcelo Cortes's parents. Their son was 11 and the commands on its screen were in English, a language he barely spoke. This might have been a discouragement. Cortes treated it as a menu. He ran every command he could find simply to learn what happened.
Before long, he was the family's resident computer expert. The title was informal, the laboratory had late fees, and the customers came in looking for films rather than software. Yet the shop gave him two educations at once. The computer taught him that an unfamiliar system could be understood through patient experiments. His parents taught him that a small business is a daily sequence of bets: what to stock, whom to trust, where to spend, and what to do when the customer wants something you failed to anticipate.
That pairing of code and commerce would become the long rhyme in Cortes's career. He moved to Canada to study computer science at the University of Waterloo, worked on products at Google, helped turn Square Cash from a prototype into a large-scale service, then co-founded Faire, a wholesale marketplace built for independent retailers. A boy learned computing inside a shop. The engineer eventually built computing for shops. Careers seldom arrange their symmetry so politely.
One problem, carried across four stations
Intel 286
A new home
Cash App
Less risky buying
The education of a useful machine
Cortes attended a Brazilian high school that offered information technology courses, which deepened his interest in programming. His father had imagined mechanical engineering for him. Software won. At Waterloo, Cortes entered an ecosystem with an unusual front door to global technology companies: serious technical training, co-operative work and a habit of sending graduates into ambitious engineering organizations without requiring them to abandon Canada forever.
At Google, where he became a senior software engineer, he worked on Hangouts, Google+ and other large products. At Square, he served as a staff engineer on the service then known as Square Cash. By 2019, a University of Waterloo biography credited him with helping take it from prototype to a product downloaded more than 30 million times. Square also supplied the future Faire founding team. Cortes met Max Rhodes, Jeff Kolovson and Daniele Perito there. Their later venture did not emerge from strangers discovering one another over artisanal coffee. It came from colleagues who had already watched each other work.
Faire launched with Y Combinator's Winter 2017 cohort. Its proposition addressed a peculiarly physical problem. Independent retailers must buy goods before they know whether shoppers will want them. Every charming object on a shelf is also cash waiting, with varying degrees of patience, to become cash again. Large chains have purchasing departments, datasets and bargaining power. The owner of a neighborhood shop has taste, local knowledge and a finite bank balance.
Faire brought brands, retailers, payment terms, returns and product discovery into one marketplace. Cortes became its technical leader and built the engineering organization alongside the product. He also put a Canadian flag on the map. Faire established its engineering headquarters in Kitchener-Waterloo and expanded into Toronto, then distributed its technical teams across New York, San Francisco, London and São Paulo. By 2025, Faire said the organization had grown beyond 300 engineers.
Canada, with ambition imported carefully
Cortes has become an advocate for building global engineering companies in Canada, though not for turning the country into a tribute act. He speaks about carrying some of Silicon Valley's operating DNA north: urgency, knowledge about scale and the conviction that a company built here can serve the world. Canada contributes different advantages, including quality of life, technical depth and employees who are less inclined to change jobs with every passing funding round.
His biography makes the argument personal. Born and raised in Brazil, he found a home in Waterloo, connected that home to Google and Square, and then used Faire to pull more opportunities back toward it. In 2025 he told the BetaKit Podcast, “I think every ecosystem is different, but we can learn a lot from Silicon Valley, and we can merge what we have here.” The verb matters. A merger keeps both parties visible. A copy leaves only the toner.
Brazil remained part of Faire's engineering plan as well. In the year before Cortes appeared at Web Summit Rio in May 2023, the company's team in the country grew by 600 percent to nearly 40 people. The expansion joined the places that shaped him instead of asking him to select a single one. São Paulo supplied the early curiosity, Waterloo the technical formation, Silicon Valley the scale, and Faire the institution that could combine all three.
When the recommendation was too good
Retail appears to be a natural target for artificial intelligence because it produces a wealth of signals: searches, orders, reorders and returns. But Cortes's most revealing AI story begins with a product that did not work. Faire once experimented with using data to hand retailers a specific list of products to buy. The response was cool. An independent shop is an argument about the world, made with ceramics, cookbooks, socks and scented candles. Owners were not eager to outsource the argument.
The team changed its approach. Rather than presenting an algorithmic shopping list, Faire used its systems to surface relevant products inside discovery and left the act of merchandising to the retailer. It was a correction with philosophical consequences. Data could widen the field. It could lower risk. It could notice a promising match across thousands of brands. It should not turn an owner into the caretaker of a machine's taste.
At HumanX in 2026, Cortes described a concrete result of improved matching. Faire's return rates, he said, were around 30 percent when the company started and had reached the single digits. A lower return rate is useful accounting. It is also evidence of confidence: the owner can try something new with better odds, the brand finds a shop where its work belongs, and fewer boxes make the melancholy trip back to the warehouse.
Faire's tools have grown from early experiments with ChatGPT in search to natural-language queries and image-based discovery. The most persuasive version of this future is almost modest. AI works backstage, searching a larger catalogue, reading patterns that no proprietor has time to calculate, and freeing the person at the counter to do the social and aesthetic work that made the store worth entering.
The architect returns to the drawing board
In 2024, Cortes gave up the CTO title and became Faire's Chief Architect when former Uber and Coupang CTO Thuan Pham joined the company. Founder stories often treat a title as a trophy that must be defended from dust. Cortes framed the change as a chance to work directly on new technology while remaining on the executive team. After years spent establishing the organization, the move put him nearer the systems it existed to build.
The timing suited his interests. AI was moving from presentation slides into the ordinary machinery of search, merchandising and advertising. In 2025, he helped introduce Faire's Promoted Listings to Canadian brands, using automatic targeting to connect products with retailers likely to order and reorder them. He also joined Toronto Tech Week conversations about engineering culture and, later that year, Maropost's advisory board.
Yet the durable idea in Cortes's work is older than any model. Technology earns its place when it returns agency to the person using it. His parents' computer let an inquisitive child probe a new world one command at a time. Faire aims to let a retailer probe a market without risking the entire shop on every guess. Between those scenes lies a career in increasingly complicated software and a remarkably consistent instinct.
The independent shop will still need a point of view. It will still need an owner willing to decide that one peculiar vase deserves the window. Cortes is betting that software can make the wager less punishing while leaving the nerve, the eye and the story intact. Even an intelligent machine, after all, should know when the customer has better taste.