From the start of the pilot project to the finished use case: as an analytics consultant for advanced analytics and artificial intelligence (AI), Dr. Christiane Glatz is in action wherever problems are to be solved with the help of AI. In this interview she tells us why moving from academic research to Daimler AG was an easy decision, and how she is bringing AI forward for Daimler with her inter-disciplinary work.
Ms. Glatz, you originally came from the research sector. What contacts did you have with Daimler before you joined the corporation?
That is correct. As part of my research, I placed my focus on HMI, or human-machine interaction. This also included attending specialist conferences, where I first made contact with Daimler AG personnel from Research & Development. When we got into conversation, I found that there were major overlaps in our knowledge and activities. So I thought to myself: that is a department I'd like to be part of!
What do you mean exactly?
During my doctorate in the field of cognitive and neuro-science, I carried out a great deal of data registration and evaluation, especially for the analysis of brain signals in the context of in-vehicle warning signals. Basically, I was working with data in a way similar to my Daimler colleagues in Advanced Analytics. I had already been planning to move from science to industry for some time, and to transfer my knowledge. And when I found a position that was precisely in my field, my decision was made. Now I'm here!
What's your area of activity now?
I'm working as a consultant for advanced analytics and artificial intelligence in Daimler's in-house "Center of Excellence Advanced Analytics and Big Data". My job is very varied: the focus of our data science teams is of course on specific problems in our corporate units, and on the extent to which we can provide a "digital value", i.e. an added value, with big data and machine learning. In addition, we acquaint our colleagues with the subject of AI with workshops or talks.
So I build a bridge between the knowledge in our department and the business cases of the individual business units. The key question is always this: can the problem be resolved effectively by AI at all? After all, AI doesn't mean that we can "conjure up" predictions. The prerequisite for successful projects always is the available data.
Switching from research to a business corporation must be a major change. You have now been with us for around one year, how do you like it so far?
I like it a lot! Daimler is very advanced when it comes to AI. There is an outstanding exchange of information about scientific topics and methods via various networks. This is great, especially if you come from a research background. Moreover, I've had nothing but positive experiences in our team. We are multi-disciplinary as a department, and greatly benefit from exchanging information amongst ourselves. We have political scientists, physicists and specialists in business informatics working together as data scientists. The projects I have worked on so far have been exciting and varied. It motivates me that my working results don't lie dormant, but generate an added value for the company in the longer term.
Can you give us one of your project highlights to date?
Last year I was in Canada with eXtollo, Daimler's in-house, Microsoft Azure-based big data platform, where I worked on a use case with our local colleagues. We showed them how to record and analyze their data with the help of eXtollo in future. It's also wonderful to find that your own work creates an added value, and that use cases are actually implemented; at the same time it increases the motivation of the specialist departments when they see what can be achieved with AI.
Is Daimler on the right track with respect to AI?
Based on my experiences so far, I would definitely say "yes". AI models naturally need well-prepared data to be successful. The volume of data we generate and process is constantly increasing, and the Group's central database – the eXtollo Data Lake – also increases our possibilities.
The corporation is internally networked with talks and exchange platforms, but we also take part in external conferences to learn more. We also cooperate with universities on an ongoing basis. The rapid growth of this research area doesn't allow you to isolate yourself.
What attributes do you need to work in Data Science?
Apart from an enthusiasm for data: receptiveness to new things! I am now completely immersed in IT, and had to learn quite a bit that was new, as I didn't study computer science. As a data scientist you must naturally be able to program, and I do this with in Python for example. As a consultant, the focus is rather more on customer contact, including the transfer of knowledge and solutions. Of course, we gladly support each other!
One last question: How would you describe your current activities to a child?
[laughs] I would say that I analyze a great deal of information to make predictions. I look at what happened in the past, and for signs that lead us to suspect that the same might happen again. I give this information to the computer, and the computer creates a predictive model so that others can also benefit from this knowledge.