Real-time Train Staffing Challenge: Maximize Reliability for Customers

We are looking for... 

…process mining, machine learning and AI-specialists to develop a machine learning approach in order to support real-time decision making for train crew staffing when unforeseen things are going on in our network – and trust us, we deal with these every day!

The vision

Help us to optimize real-time train staffing in 3 steps:

Step 1: Learn the rules
Find patterns in historic shift plan data for train crews

Step 2: Identify conflicts
that could arise in workforce scheduling due to unforeseen delays

Step 3: Support live-action!
Develop a software to analyze where planning interventions might be necessary

Recognize conflicts in workforce scheduling due to unforeseen events. Picture this: because of a bottleneck in the system staff changeover cannot take place as planned. If e.g. maximum shift length of the train crew is exceeded or legally required breaks cannot take place as planned a train might be left without crew – what will passengers say?

In order to take the right decisions quickly we would like to anticipate conflicts as early as possible: When does it make sense to counteract this in workforce planning? At which train stations should a staff changeover take place or where is which staff so spontaneously available?

Join our challenge! The Steps:

5-day-challenge (virtually): EUR 3k for 5 participants

Follow up for Challenge winner: 100 days PoC- program

Long-term-success: chance of implementation at Deutsche Bahn


All applications received will be evaluated by DB and 5 Teams will be selected for participation in the Challenge week (5 days). They have the chance to develop a mockup of their solution– maybe for a specific part of the problem – in close sparring with topic experts (daily exchange sessions) and guided by DB mindbox team members.

On the last day a virtual pitch event will be held. The participating teams will present their live simulation to a jury of DB employees and a winner will be picked. The winner will instantly kick off a 100 days program to develop their mockup into a prototype that includes the capacity to solve the identified conflicts (see timeline below).


During the challenge week, we expect a continuous virtual availability via MS Teams to enable an intensive exchange between the participating teams and the specialist staff of DB.


Each participating team receives a one-off payment of 3.000 € for their total commitment to our Challenge during the 5 days.

Who can apply?

All process lovers, data experts and teams with thorough expertise in process mining, machine learning and AI-architecture and a love for optimizing train operations by smart and predictive workforce planning.

This challenge is an open contest for all creative founders as well as established companies.


Application Deadline
Invitations to Challenge participants
Data Briefing Call
Challenge Week (Top 5 Teams)
06.07.2020 - 10.07.2020
100 Days Program for Challenge winner