The journey towards a data-driven company

In today’s digital era, data offers companies enormous added value. Audi uses this potential to make desicions based on data, establish new digital business models and improve core processes. As Audi’s Data Team, we therefore focus on introducing structure where there is none and shaping tomorrow’s products from today’s data. We also ensure a consistent Data2Value journey.

Thinking big, initiating new ideas – that’s how we work

In our team, ‘we’ is more important than ‘me’. In the division of Data Analytics / Machine Learning, our objective is to jointly create synergies in order to provide flexible and cost-efficient solutions for innovative data services on a high level of in-house development. Always on the look-out for new technologies, we lay the foundation for data projects and promote data awareness through our company-wide data community. In our international team, we work dynamically and in close collaboration with all business units at AUDI AG  as well as with the VW Group. Thanks to a culture of open feedback, we learn from each other and encourage each other in our individual abilities. We strive to make Audi even more of a data-driven company. What motivates us? We want to harness data-based knowledge and, by doing so, design tomorrow’s future together and shape it in a sustainable way.

Data at Audi

Increasing digitalisation means that we, as the Data Team, must react flexibly to changes. Both the ways in which we work together and our technological focus areas are constantly being adjusted to meet requirements. At the same time, we are able to map end-to-end chains as well as provide analysis, insights and answers in an agile way. This requires a high level of in-house development and a profound knowledge of all data domains. Based on this, we implement entire products in our own development teams. Our product portfolio ranges from traditional business intelligence to data engineering and machine learning. The matrix organisation within Audi IT and with our partners from the VW Group facilitates networked collaboration and, at the same time, embeds data awareness in our corporate culture. To drive forward data at Audi, we are organised in a competence-based way:

On the road to a data-driven company a high level of in-house development is required.

Product Management and Analysis

Product Management and Analysis

This team carries overall responsibility for the process of implementing and managing data products. This includes the design, realisation and delivery of integrated data applications for all business divisions. Further fields of activity are the analysis of requirements for data products and services, getting the divisions on board with a data development that is useful across the entire business, and the standardised use of architectures and tools. Besides our technological know-how, we also use our knowledge of the business processes, the involved data and how they integrate.

Data Engineering/Data Science and Machine Learning

Data Engineering/Data Science and Machine Learning

These internal development teams drive the integration and analysis of data sources, including solution architecture. Data Engineering is concerned with the modelling and provision of data. The Data Science and Machine Learning teams then rely on this data to gain new insights with statistical evaluations and analytical models. While the Data Science team focuses on the processing of structured data, the Machine Learning team primarily trains models for process automation on unstructured data, such as images, texts and audio data.

Solution Architecture/Data Security

Solution Architecture/Data Security

This team deals with topics related to information architecture and technical architecture of data applications and platforms. Together as a team, we continuously develop the technical portfolio. Further aspects are advising data experts from all business areas and the joint conception of specific solution architectures for ‘on-prem’ and cloud – depending on the application, we even deliver a hybrid analytical architecture. Moreover, standards are defined in order to ensure that developments on our data platforms comply with governance and IT security guidelines. The team is a multiplier for IT security standards and supports Audi IT’s central approval process.

Data Analytics/Machine Learning Platform

Data Analytics/Machine Learning Platform

Our data platforms present a centrally orchestrated technology stack as a flexible and integrated solution for all business divisions at Audi as well as for the VW Group. We work together intensively with many internal and external partners. With the jointly designed portfolio, we are able to guarantee a quick and agile development of end-to-end products. Our development strategy also allows the achievement of sustainable standards for data and tools.

Insights into the Audi working world

Do you speak ‘Vorsprung’? Take the test!

With the rise of data, we also need methods that enable us to generate knowledge from data. Our colleagues in the Machine Learning Team are specialists in the field of deep learning. They use unstructured data, such as images, text and audio data, to work more efficiently in a wide range of areas – from production and finance to design.

Matthias Graunitz and Mike Winkelmann explain in ‘Audi speak’ how machine learning helps us to become more efficient and why a hybrid architecture is crucial. If it doesn’t sound like double Dutch to you, then you’ll fit right in!

What we define is used in many business divisions – both at AUDI AG and at the VW Group.

Nicole Burtz, front-end coordinator on the data analytics/machine learning platform

Shaping tomorrow’s products from today’s data

Nicole Burtz and Thomas Hager deal with complex topics related to machine learning and data analysis in their area at Audi. They work on different projects, but as a common unit for cross-divisional data issues in the whole company. Their aim: to make Audi even more of a data-driven company. The benefit for customers: a better product experience thanks to a tailor-made service. A conversation about hairline cracks, engine noise and mariachi bands.

Analysing data today to learn what will be important tomorrow: that, in essence, is the job of the Data Analytics/Machine Learning organisational unit. Whenever there is a need to manage company processes based on data, to improve and enable new business models and to bring about the best possible decisions, it is the team’s turn.

With its analysis and its sustainable data management, the division supports the implementation of projects – and ensures a successful data-driven business. We spoke to two team members who are fully committed to this goal.

More details

Nicole Burtz, you took your A levels by distance learning almost as an after-thought, as you were already working at Audi. What support did you receive from your employer?
Burtz:
At that time, I had initially started vocational training to be an IT Business Administrator at Audi and then decided to catch up on my A levels. It was great that I was able to take study leave to prepare for my exams. I thought that was a fantastic concession. As was the guarantee of re-employment that Audi then gave me, which meant that I was immediately able to study full time after my A levels. The company has therefore supported me on a very long road – and across many stations in my career.

Thomas Hager, you’re a fan of Ingolstadt FC – how could modern IT applications, such as those you use at Audi, help the club to avoid relegation?
Hager
: No big data app in the world can guarantee the results you want! But, generally, detailed data analysis are no longer inconceivable in football and can serve the coaching team as the basis for tactical decisions. We have various tools for this at Audi. However, in our case the evaluation is much more complex because our ‘pitch’ is a much bigger one, and the rules change must faster!

Thomas Hager, project manager, BI products/analytics projects

Thomas Hager, project manager, BI products/analytics projects


What type of machine learning do you recommend: a deterministic approach, i.e. as ‘supervised learning’ or in the sense of a deep learning with iterative processes?
Hager
: In my view, machine learning is of particular interest when you can make predictions for the future from historical data. It serves employees as a decision-making tool for complex issues. At Audi, for example, we are currently training an algorithm that analyse engine noise and – based on that – is able to conclude which component is faulty.

Burtz:
 Furthermore, we use deep learning for crack detection. Using camera images, we try to spot fine hairline cracks with the help of AI. This, in turn, helps to detect defects in the material at an early stage and to provide an even higher quality in our products.

More details

“Due to the data gravity, we have chosen a hybrid platform approach, which we continue to expand.”

Nicole Burtz, front-end coordinator on the data analytics/machine learning platform

How does the process of collecting and analysing data look like?
Hager:
With the active consent of the customer in the vehicle, we are now gathering millions of pieces of data in anonymised form from our vehicle fleet. That’s why we have a whole team here dedicated solely to analysing this data with a wide range of tools, always with the objective of better understanding our customers’ wishes and continuously improving our products.

You are using Hadoop to process, or rather parallelise, large amounts of data. Some components of this (HDFS) are considered outdated because of high latency – how are you getting around this at Audi? The keyword here is:
cloud computing.
Burtz:
As times change, we constantly undertake tech scouting, with the aim that we are always providing a secure and stable service. We buy technologies like Hadoop, then build our own applications on top of them, thereby providing multi-project platforms for Audi and the Group. Changes to these applications can also lead to a change of platforms and technologies. Through a Proof of Technology, we carry out a feasibility analysis within the Audi infrastructure and can quickly make decisions independently in the team using in-house development. In this way we try to offer our internal customers from the business departments the best possible portfolio.

In future, will applications all be cloud-based or will on-premise solutions remain just as important?
Burtz:
We won’t rely solely on cloud solutions any time soon. Some data is simply too critical to be stored in a public cloud. At Audi, we’re bound by our own code of data ethics, i.e. the protection of sensitive data. Data gravity must also be taken into account. This means that the place where the data is created is crucial for where we store the data. We've therefore deciced to go with a hybrid platform approach and keep on expanding this.

Nicole Burtz, front-end coordinator on the data analytics/machine learning platform

Nicole Burtz, front-end coordinator on the data analytics/machine learning platform

Is this idea of responsible data ethics behind the set up of the Hybrid Audi Analytics Platform (HAAP)?
Hager:
Precisely. HAAP is our central data analytics/machine learning platform and serves in my current project, for example, as a reliable and secure source for extracting data and incorporating it into our multi-dimensional planning app. Specifically, the project is about setting the course as early as possible in a product development process. This means: which products should come on to the market when and with which body work and which drive system? These are very significant strategic decisions for which we need planning and simulation tools. The software is very important here as it enables us to get a great deal right at an early stage.

Burtz:
With HAAP we have, on the one hand, the data storage level in which we locate back-end components – traditional business and data warehouses. And then we have systems which we build on to these to run data analysis – there are various front-end components. So, we try to operate a large number of use cases, ranging from traditional BI (business intelligence), planning and simulation to advanced analytics.

More details

“We rely on agile project management. It allows us to react much faster than with the traditional ‘waterfall’ project’.”

Thomas Hager, project manager, BI products/analytics projects

How important is project management for such critical decisions, as well as in the team?
Hager:
These days we carry out almost all projects using agile project management – the keyword here being ‘scrum’. This is firmly embedded in our day-to-day work. Consequently, we have very intensive team collaboration and, at the same time, short communication channels. Our ‘project dailies’, in which we manage to exchange information with the whole project team in just 15 minutes, are an example of this. Where are the obstacles to progress? Where do we need to react? This allows us to identify challenges at an early stage and address them quickly. It’s in contrast to the traditional ‘waterfall’ project management approach in which function and requirement specifications first have to be written down and then ‘tossed over the fence’ to the development team for a solution to come back six months later – those days are gone. We now move forward in agile sprints of two to four weeks.

Burtz:
I can only underline what Thomas says. We have switched our entire platform development to agile working, which has helped us to create transparency – and that’s the essential thing for me. We can identify our project blockers much faster thanks to our agile approach. Our teams are based all around the world, be it in India, Poland or Germany. Nevertheless, thanks to collaboration tools, we can coordinate closely and promptly.

“I can’t think of a more exciting area at Audi at the moment where I’d rather work.”

Nicole Burtz, front-end coordinator on the data analytics/machine learning platform

What is the biggest challenge you face in implementing these complex projects?
Burtz:
The biggest challenge is bringing everyone together to build these services for the company as a flexible multi-project platform, without creating siloed applications. What we define is used in many business divisions – both at AUDI AG and at the VW Group.

Speaking of challenges: Thomas Hager, you’ve also worked for Audi Mexico – what positive influences did you bring back with you for your work here?
Hager:
In Mexico, creating a good personal relationship is crucial for a good professional collaboration. I was offered something special almost every free minute, from sight-seeing to mariachi bands. We could learn a thing or two from this hospitality here in Germany. At least since then I’ve been trying to “be more Mexican” with my own team. It fits in well with Audi’s lively working atmosphere.

Nicole Burtz, have you experienced a ‘Mexican’ working atmosphere too?
Burtz:
I certainly haven’t seen any mariachi bands yet, but I have a really nice international team around me who I learn from every day and who I love working with. I can’t think of a more exciting area at Audi at the moment where I’d rather work!

Close

Audi worldwide

Models, products and services – switch to your country / sales region website and discover the regional diversity of Audi.

    Close

    Close

    You are using an unsupported browser to access this website. To get the best User-Experience while visiting the website, please use the latest versions of Chrome, Firefox, Safari or Edge.

    Thank you very much for your visit.
    Your Audi Team

    The international Audi website

    Discover Audi as a brand, company and employer on our international website. Experience our vision of mobility and let yourself be inspired.

    Audi of America: models, products and services

    Explore the full lineup of SUVs, sedans, e-tron models & more. Build your own, search inventory and explore current special offers.

    Switch to audiusa.com

    Error report

    The feedback form is currently unavailable.
    Please try again later.

    You may deactivate your ad blocker to view the feedback form.