Introduction
Armin Yilmaz can name streets where navigation systems behave as if a city were only geometry. They see congestion, but not carnival days, unofficial taxi habits or the mood after a concert. Thirty years behind the wheel teach a different kind of map.
Armin lives in Cologne and now works with a local mobility lab. He brings driver knowledge into AI projects that would otherwise treat movement as lines on a screen.
Story of the Path into AI
Armin began with irritation. Apps suggested routes that made sense in general and failed in detail. A taxi rank that looks empty in the data may be useless after midnight; a quiet street may become important after a specific venue closes.
At tech meetups he initially felt out of place. Younger developers spoke about mobility without ever having driven a night shift. Armin learned geodata, simple statistics and privacy rules for movement data. His first project was a map combining waiting times at taxi stands with weather and events.
The first version missed informal knowledge: which exits people actually use, which police barriers appear during events, which shortcuts passengers dislike. Armin added driver interviews to the evaluation.
Current Work
Today Armin helps test demand forecasts for urban transport. In one model, a seemingly quiet street segment was rated irrelevant. He explained that after certain concerts it becomes the quickest way to avoid a blocked square. The forecast changed after local knowledge was included.
The point is not to romanticize experience. Drivers can be wrong too. But when AI systems shape mobility, they need the knowledge of people who live inside the rhythms, not only the data traces left behind.
Personal Advice
“Data shows movements. People often know the reasons behind them,” Armin says. He advises transport teams to ask workers what the map hides before optimizing routes.
Key Facts
Age and place: 54, Cologne.
Background: taxi trade, shift work, informal city knowledge.
Entry into AI: map linking taxi-stand waiting times with weather and events.
Focus today: urban AI and mobility planning.
Typical tools: geodata, demand forecasts, mobility analysis.
Werkstattnotiz
Armin’s notebook contains street names with arrows, jokes and warnings that would look unprofessional in a report. Some of them explain forecast errors better than the clean dashboard. He is now working on a way to include driver notes without turning people’s routes into surveillance.