He started selling online at 17. Seven years later, the CEO of Pango is building an AI agent to run the deliveries, tracking, and returns that most founders would rather forget.
Most people who build software for online stores start from the outside. They fall in love with the storefront, the checkout, the moment a shopper clicks buy. Steve Rahimi went the other way. He walked into the back of the shop, the part with the taped-together spreadsheets and the seven browser tabs, and decided that was where the interesting work lived.
Rahimi is the CEO and co-founder of Pango, a Stockholm startup in Y Combinator's Summer 2026 batch. The company describes itself as an agentic operating system for e-commerce operations, which is a technical way of saying it wants to be the digital employee that runs everything after a customer hits purchase. Deliveries. Tracking. Returns. Refunds. The customer service message that arrives at 2 a.m. asking where a package is.
It is not a glamorous problem. That is precisely the point.
Rahimi has said he became an e-commerce entrepreneur at 17. That detail matters more than a founder-origin flourish usually does, because it explains the shape of everything Pango is now. He did not study the problem from a business school case or a consulting deck. He lived inside it, one order at a time, back when the job of running an online store meant doing all the unglamorous parts yourself.
By the time he was building Pango, roughly seven years of that hands-on operating had accumulated into a very specific kind of knowledge. Not the theory of e-commerce. The texture of it. Which carrier goes quiet when a package is late. How a return actually moves through a warehouse. Why a customer who cannot track their order becomes a customer who never comes back.
That line, from a piece he wrote on choosing returns software, is a small masterclass in reframing a category. A returns portal is a webpage. An operation is a workflow, with carriers and warehouses and refunds and exceptions all firing at once. Rahimi's argument is that brands keep shopping for prettier webpages when what they actually need is a system that runs the whole messy chain. It is the sort of observation you can only make after you have been the person doing the running.
Before writing much product, Rahimi and his co-founder did something more operators should do and fewer founders actually will: they embedded with brands. Around thirty of them. What they found was a familiar kind of chaos. To run the back office of a scaling store, teams were stitching together five to seven separate tools that did not talk to each other. One for the return portal, one for tracking, one for carrier rates, one for the customer service inbox, and a person in the middle copying data between all of them.
Fig. 1 - The pitch in one diagram: collapse the tool sprawl into a single agent that runs it.
Pango's answer is to let a merchant describe how they want their operation to run, in plain terms, and then have agents build and run it. The system assembles return portals, optimizes delivery options at checkout, orchestrates multiple carriers, and deploys agents to handle the customer service tickets that pile up around logistics. The company says it automates about 99% of returns and shipping operations while trimming operating costs by around a fifth.
Rahimi did not build alone. His co-founder and CTO is Lukasz Reszczynski, whose resume reads like a different genre entirely. Reszczynski holds a master's in astrophysics and did doctoral work in molecular biology before turning to software, where he built a configurator that reportedly powered more than $1.5 billion in product quotes. The pairing is almost too neat: the operator who lived inside the problem, and the scientist who likes building systems that hold up under load.
It is a common pattern in the strongest founding teams, the domain expert next to the deep builder, and it tends to work because each one keeps the other honest. Rahimi brings the ground truth of what actually breaks in a store's back office. Reszczynski brings the machinery to automate it without the whole thing falling over the first time a carrier API changes.
The early numbers are the kind that make investors sit up. Pango says it reached product-market fit roughly two months after launch, then grew about 60% month-over-month heading into Y Combinator. Thirty-plus brands across fashion, beauty, and home went live on the platform. The company took Swedish innovation grants early, incorporated as a US C-Corp to chase American growth, and raised an angel round to keep the flywheel turning.
Fig. 2 - Compounding at ~60% a month bends fast. Shape is illustrative of the reported growth rate.
What is striking is not any single metric but how quickly the pieces lined up. Fast product-market fit usually is not luck. It is the reward for choosing a problem you already understand at the level of muscle memory. Rahimi had spent years absorbing the frustration of post-purchase operations before he ever tried to sell the cure, which is why the cure landed so fast.
Ask most brands about logistics and returns and they will describe a cost center, a line item to be minimized and endured. Rahimi's larger ambition is to flip that framing. If the post-purchase journey is run well, a delivery becomes a reason to buy again and a smooth return becomes a reason to trust the brand. The friction that quietly drains loyalty can, handled right, become the thing that builds it.
That is the real wager underneath the agent demos and the automation stats. Pango is not just trying to make the boring parts cheaper. It is betting that the boring parts, once they run themselves, become a competitive advantage that scaling brands cannot easily copy. For a founder who went looking for the ugliest job in online retail, it is a fitting place to plant a flag.
Rahimi is still early. Pango is a small team pointed at a very large, very unglamorous market, and the story of whether agents can truly run a store's operations is only starting to be written. But the setup is the kind that tends to age well: a founder who knows the problem in his bones, a builder who can hold the system together, and a category everyone else finds too dull to fight over.