The cleverest thing about Vadym Melnyk's drone is not that it flies. Flight is the showy part, a trick birds mastered before venture capital had a name. The serious part begins when the battery fades. The aircraft returns to a box, settles onto its mark, exchanges or replenishes its power, and prepares to work again. No engineer jogs across a refinery carrying a spare. No applause is required. The machine simply comes home.
That return trip is a useful way to understand Melnyk. His career has crossed countries, institutions and fashionable technologies, but it keeps landing on the same practical square: take a task that is dangerous, repetitive or slow, and build a system that can repeat it reliably. Dronehub applied the idea to industrial inspection. VADYM.AI and KIERUNEK.AI applied it to office work. Oswin AI, his new American venture, brings him back to the border between software and machines.
He was born in 1994 in Ivano-Frankivsk, Ukraine. As a teenager, he won a Future Leaders Exchange scholarship and spent a school year in the United States with an American family. The experience did not hand him a company, though founder biographies often make geography look like destiny with better airport lounges. It did widen the map. Poland came next, and the University of Information Technology and Management in Rzeszów.
Before the drone, an intelligent guitar
Melnyk had been competing in robotics while still young. At university, the projects arrived with the cheerful disorder of a curious engineering mind. One was an intelligent guitar designed to teach its user to play. His girlfriend learned. He did not. He later delivered the joke himself: proof, at least, that the device worked.
Drones entered during his second year, while he was studying in Spain. He and three others formed a team. In 2015, their autonomous aircraft and cloud system for spotting threats earned third place in Microsoft's Imagine Cup national finals. Microsoft became more than a logo on a competition slide. As a student partner, Melnyk spoke at conferences, learned new tools and acquired the difficult soft machinery of entrepreneurship: a network, an audience and a reason for strangers to return an email.
“We built our name, our credibility, thanks to Microsoft.”Vadym Melnyk, recalling his student-partner years
He eventually became Chief of Microsoft's Unicorn Division, a startup-support program he helped create for Central and Eastern Europe. The title was not imposed by a committee. Melnyk negotiated for it because he wanted “Unicorn Division” on his business card. Beneath the gag was useful work: scouting software companies, connecting founders with venture capital and helping them reach Microsoft's cloud programs.
A company finds its problem
Cervi Robotics began in 2015, during Melnyk's final year at university. He left before graduating to build it. The young company was energetic and indiscriminate: custom drones for defense and police, factory automation, Industry 4.0 projects, cloud computing, data work. It earned revenue, hired people and risked becoming good at too many unrelated things.
Then a hard request supplied a center. In 2017, a European Space Agency contract asked the team to develop a drone that could land without an operator and change its own battery. Cervi produced a working prototype that year. By 2019, supported by grants, private investment and money from its services business, the team had a product it could put in the field.
The route from experiment to infrastructure
The first commercial customer was PKN Orlen. At its Płock refinery, the system completed 26 patrol missions, testing the unromantic details upon which industrial technology lives or dies: repeated flights, real infrastructure, useful images, weather, schedules and the expectation that tomorrow's mission should resemble today's. A pitch deck can omit drizzle. A docking station cannot.
The product joined three things that are often discussed separately: the aircraft, its ground infrastructure and software for missions and data. That integration allowed a drone to patrol fences, inspect assets or collect measurements, then return without turning every battery cycle into a human errand. The box was not furniture. It was what made autonomy operational.
The expensive value of saying no
Success presented an awkward choice. In early 2020, Cervi had nearly three million złoty in potential contracts for work outside the Dronehub product. Dronehub itself had one customer. Melnyk and his team declined the broader work, narrowed the company to autonomous drone platforms and software, and rebranded around the product.
Revenue across robotics, outsourcing and custom projects.
One customer, one hard problem and the chance to build a category.
It is tempting to dress this as fearlessness. More likely it was fear with a spreadsheet. The calculation was that a focused product solving a recurring physical problem could become larger than a busy engineering consultancy. Refusing paid work is one of the least photogenic acts in company building. Nothing launches. The inbox merely grows quieter.
The bet gained institutional weight. Dronehub worked on European research programs and industrial projects involving organizations including the European Space Agency, the European Defence Agency, Deutsche Bahn and Honeywell. Melnyk appeared on Forbes youth lists in Poland in 2020 and 2021 and in Ukraine in 2023. Dronehub appeared in the Financial Times FT1000 in 2023. Awards are not batteries, but they can keep doors powered long enough for a young industrial company to enter.
America, again
In 2022, Dronehub joined the sixth GENIUS NY cohort in Syracuse. The accelerator selected it as a finalist and awarded $500,000 to support U.S. growth. The company opened an American operation, while Melnyk later established himself in North Carolina's Research Triangle. His old exchange-country had become a market and a base.
The geography matters because hardware companies grow through ecosystems as much as slogans. Rzeszów offered an academic pre-incubator and Poland's Aviation Valley. European programs gave the team difficult briefs and research partners. Syracuse supplied a drone accelerator and an entry point. North Carolina adds universities, aerospace activity and a dense technology corridor. Melnyk's career reads less like a straight line than a set of well-chosen docking stations.
He has also resisted becoming only the drone founder. VADYM.AI in Ukrainian and KIERUNEK.AI in Polish teach entrepreneurs to build practical automation. By his account, the programs have reached tens of thousands. His teaching emphasizes the gap between seeing a demo and changing a workflow. A perfect demonstration can make a tool look magical. Tuesday afternoon, with untidy company data and a wrong first output, is where skill begins.
“If I do something twice, I think about automating it; if three times, I automate it.”Vadym Melnyk
It is a small rule with the shape of his whole career. A refinery patrol repeated often enough becomes a drone mission. A drone mission repeated often enough requires a dock. An office chore repeated often enough becomes an AI workflow. The object changes; the irritation with avoidable repetition remains.
Failure, with telemetry
Hardware gave Melnyk a sturdy philosophy for the AI era. Before a Dronehub aircraft could land reliably, prototypes failed. Each bad approach returned information: alignment, control, weather, tolerances, another variable refusing to behave politely. He describes those failures as the product's raw material. An unsuccessful landing is not inspiring, but it is specific. Specific failure can be repaired.
That is why his AI advice prizes shippers over spectators. People who stop after a poor first output treat failure as a verdict. People who adjust the prompt, data or workflow treat it as telemetry. The distinction is more durable than any model release. It also explains his trajectory from the physical patience of robotics to the faster iteration of software automation.
In 2026, he added Oswin AI, a U.S.-based company working at the meeting point of AI and robotics. Public detail remains spare, but the direction is consistent. He is interested in autonomous systems, applied AI, dual-use technology and the movement of deep-tech companies between Europe and America. He also presents himself as an adviser and investor for founders entering those markets.
The ambition is broad now, yet the best clue remains that box on the ground. Flight attracts the eye. The dock creates the routine. Melnyk's work is an argument that autonomy is not a spectacular act performed once; it is a dependable loop. Leave, observe, return, learn, repeat.
There is something pleasantly unfuturistic about that future. It consists of fewer dangerous inspections, fewer repeated emails and fewer people dispatched to rescue a supposedly autonomous machine. The robot earns its intelligence by being useful. Then, when the work is done or the battery is low, it has the good sense to come home.