Konux
AI monitoring for safer, more reliable railways
Konux is a German AI scale-up that brings smart monitoring to railways. Its sensors and machine learning track the condition of the tracks in real time and turn the data into clear insights, so operators can predict wear, plan maintenance, and keep trains running on time.
My role
As Product Designer, I tackled design problems in the railway industry across the German, UK, Spanish, and Japanese markets. I designed clear UI for data analysis, visualized complex technical processes, and shaped UX that helps operators and engineers decide faster.
Challenges
01
Engineers often run maintenance and checks that are not needed, which leads to budget overruns, delays, and wasted resources. Without precise monitoring, real issues get missed and unplanned work disrupts traffic.
02
There is a huge volume of data, and going through it takes time. It is valuable but lacks clear priorities and practical advice, so engineers struggle to act on the most urgent issues first.
03
Teams often react to faults only after they appear. Without early warning of wear, small problems grow into failures that cause delays and expensive emergency repairs.
Holistic view
A track node has several parts, like the points, frog, and point machine. To decide well, engineers need both the overall health of each part and detailed values like vertical displacement and wheel path. Our device captures all of this. My task was to present that data clearly and keep it scalable as the volume grows each year.
First iteration
After several rounds with stakeholders and users, usability testing showed this version worked best. It shows detailed data for every part while highlighting the most critical facts about the node, so engineers see both the detail and the big picture. It also scales as more data comes in.
Insights instead of raw data
In interviews with engineers from several countries, I saw the same pattern: they lacked insights. They rarely have time to study raw data and prefer clear instructions on what to do and where. So we worked with data engineers on per-component recommendations. Each one is data-driven and shows a clear priority, so engineers know which part to fix first.
Mobile App
So inspection engineers can log issues on the spot, we built an MVP mobile app. All data goes to a central database and syncs automatically with the client's system.
Traffic
A client came to us with a critical problem: one train just five minutes late could push others 10 to 15 minutes behind. These knock-on delays can spread across the whole network and even bring it to a stop. Users needed to trace the full chain, from the first delay to its ripple effect, so they could find root causes and prevent future disruptions.
Konux Traffic lets users follow the whole chain of events, from the first delayed train to the fix. It shows where and how often delays happen and highlights the triggers and their impact, so operators can improve punctuality.
To test it, I built several interactive prototypes, each exposing the weak spots of a different approach. The final one, refined across these rounds, went into production.
Immersion is key to great UX
To really understand the domain and the users, I went beyond desk research and into the field. With the Konux team, I helped install 16 devices right on the tracks. You can spot them if you take the train between Karlsruhe and Mannheim.