The hackathon project that made it to Ukraine: Silent Mesh

At TAIKAI, we often ask ourselves what happens to the projects built on our platform. Whether they have real utility, what path they take after the hackathon ends, and above all, who is behind them.

Last March, Paulo Fonseca took part in the EUDIS Defence Hackathon and Mentoring, a European Commission initiative that supports defense innovation among European startups and builders. He brought a new take on a simple idea, almost too obvious to be true: an acoustic drone detection system made of a mesh of fiber optic connected sensors. Eighty days later, that idea was being tested under real conditions, in an official crash test run with the Ukrainian Military Forces, out in a field around Kiyv, detecting real combat drones.

Between those two moments, Paulo relocated to Romania to build the prototype from scratch, opened a company while abroad, tracked down the only Mac Studio M4 Max available in Portugal to power the system, and flew to Hungary to finish assembling it days before the test. He then made his way to Ukraine by overnight train, working through air raid alerts to get the system ready, before finally seeing it detect and track real combat drones in the field. Getting there, and back, took more improvisation than anyone would expect from a hackathon project.

So, naturally, we had plenty of questions to ask him.

From crypto to defense

Paulo did not start out in defense. His background is in industrial design, and until recently his work lived in crypto. The shift began at ETH Cannes 2025, where he built Open Source Military, an early attempt to bring open collaboration into a field that rarely welcomes it. That project planted the idea that defense technology did not have to be built behind closed doors by large contractors. It could start with one person, a laptop, and a room full of strangers.

That idea followed him through several cities before he landed, almost by chance, in Lisbon.

A hackathon next door

Paulo was living in Lisbon when he learned the EUDIS Defence Hackathon was happening in the building next to his own office. He decided to join with his previous co-founder, working remotely from Romania. The parts he needed to build a working prototype arrived just an hour before the event started, and he built anyway.

The idea he brought was simple, almost stubbornly so: Russian drones increasingly rely on fiber optic guidance, a method that resists jamming and cannot be detected by conventional radio-frequency systems. Paulo's question was direct. If fiber optics make these drones nearly undetectable, why not fight fiber with fiber?

His pitch at the hackathon did not land as well as he had hoped. But one moment changed everything. A judge from Critical Software pushed back with a sharper question: what happens if the enemy plays a sound designed to throw the system off? It was a hard question, but the problem was solvable. “If this is the biggest concern about my project right now, I might be onto something” - Paulo recalls thinking to himself. Silent Mesh, an acoustic detection system built to give defenders more time to react to incoming threats, was born out of that exchange. He did not win the hackathon. He left with something more useful: a problem worth solving, and proof that people who understood the field took it seriously.

The road to Ukraine

One week after the hackathon, Paulo applied to a crash test in Ukraine, organized by Snake Island Institute, Third Army Corps and BRAVE1, to evaluate new detection technologies under real conditions. He was selected.

He flew to Romania to join his co-founder, and the two spent three weeks building and testing the prototype from the ground up. When the original test date was postponed, it turned out to be a stroke of luck. The prototype was not ready. Paulo returned to Portugal to handle bureaucracy, including registering the company, before a new test date sent him to Budapest, where he finished assembling the system, and then onward to Ukraine by train, luggage packed with enough electronics to unsettle the other passengers in his carriage.

He finished the prototype in Kyiv. With GPS use restricted for security reasons near the test site, he navigated the final stretch by following road signs alone.

When it didn't go as planned

The crash test itself did not go as expected. Weather conditions interfered with the run. Then, on the way back to Poland, came the harder blow: the prototype was confiscated at the border, by understandably over zealous Ukrainian border guards.

For a system built in three weeks by two people, with the last stretch happening in a hobbit house outside Budapest and a hotel room in Kyiv, having their only hand-built prototype be confiscated at the Ukrainian border was as close to losing everything as it gets. But that didn’t deter Paulo from continuing through with his project.

What the numbers said

The results took time to arrive. When they did, they confirmed what Paulo had felt in the field. Out of 14 participants, Silent Mesh was selected among the top 4 that were able to produce almost real time 3d coordinates of the targets. Their system produced one of the most most precise detection data of the test, though with a shorter range than some competitors, a trade-off that validates their concept since their sensors are up to 50 times cheaper than the competition.

What comes next

Paulo and his co-founder are already designing the next version of Silent Mesh. An ex-Ukrainian military member has joined the team, bringing frontline perspective to a problem that cannot be solved from a lab alone. They have Ukrainian frontline unit commanders lined up to test the new prototype and gather real validation, and they already started raising a pre-seed round.

One of the hardest problems ahead has nothing to do with detection accuracy. It is deployment. Sensors cannot realistically be placed by hand along a frontline, so the team is working on how to get the system into position without putting anyone at further risk. Reducing the cost of the sensors themselves remains one of their central goals, since a detection system is only useful if it can be deployed at scale.

Closing reflection

Asked what he took away from the experience, Paulo did not talk about the technology. He talked about hackathons.

"Hackathons are the best way to start something new," he said, pointing to the human factor and collaboration they bring in an increasingly LLM-driven world. He believes the format has value far beyond a single weekend. In his view, the best way for companies to operate would be to run one hackathon every week, giving people the space to create.

Eighty days after a hackathon next door to his office, Paulo had a working prototype in a field in Ukraine, detecting drones that most systems cannot see coming. The version of Silent Mesh that gets deployed at scale is still being built. But the idea that got it started, an obvious question nobody had quite answered yet, came from a room full of strangers with a weekend and a problem worth solving. That’s the power of hackathons.