Artificial Intelligence and Machine Learning


Artificial Intelligence and Machine Learning have long ceased being a mere gimmick and they are now key drivers for digitalisation in day-to-day life as well as the economy. We develop self-learning programmes together with our customers in our innovation lab in environments such as smart grids, predictive analytics, autonomous driving, recruitment, image recognition, social network analyses as well as intelligent data collection and data search.

Our services

Knowledge transfer

Understanding the possibilities and limitations of Artificial Intelligence and Machine Learning in their practical application.

Technology assessment

Immerse yourself in the world of deep learning, reinforcement learning, performance libraries, learning systems and much more.

Development of solutions

Develop Machine Learning systems which are embedded in your individual architecture and system landscape. Stable data pipelines supply data to the algorithms.

Innovation potential

Find and evaluate ideas and test the prototypes of your Machine Learning application cases. Use the experiences from our innovation lab.

Quality assurance

Ensure that your Machine Learning applications are secure.

Current projects

Weasel – intelligent search for employee skills

We often find ourselves confronted with the challenge of looking company-internally for an expert on a specific topic. The larger the company, the more difficult the search becomes. There are often existing databases with details on employees and their skills. However, most of the time these are hardly maintained or not at all and therefore unusable. In our Machine Learning Lab we have developed a search engine – called Weasel – which is able to use unstructured data from documents, chats and repositories. We were also able to use similarity graphs to cover skills that could not be found in primary databases and make searching easier thanks to error corrections and autocompletions. We have integrated further platforms from which Weasel collects employee skills, such as SharePoint or a Wiki, but also Stackoverflow.

Mail2Department – learning in a protected environment

We have set up a protected environment in our Machine Learning lab. This enables us to use data for the learning process without having access to the content ourselves. With the help of past e-mails we were able to train a model with a precision level of 95 percent, which automatically proposes the correct recipient of an e-mail.

Application Chatbot

We have programmed a chatbot for a career website which is able to answer various applicant questions. In order to do this, the chatbot directly accesses the content of the website. The applicants‘ questions are analysed across multiple layers regarding the set intent, analysis of the FAQs and search of the website. The chatbot can answer frequently asked questions from applicants 24/7. The chatbot shows applicants a very modern way of communication and is very well received by the applicants.

Predictive quality analytics

Internal product quality analytics require the destruction of samples and are a major financial expense. In order to reduce the number of samples required, we have developed a predictive model based on technical data collected during the production process, which models the results of the quality testing.

For this, we compared anomaly detection and semi-supervised learning models with standard regression models. We tested various Machine Learning algorithms for their prediction qualities. Subsequently, we implemented the algorithm with individual adaptations to the relevant data situation. 

Autonomous driving at MunicHMotorsport

In the FormulaStudent, an international design competition for students, a Driverless Car division has been around since the 2017 season. We have advised and supported the team of MunicHMotorsport e.V. since the beginning of the project in the development and implementation of the Machine Learning components for autonomous control.

Thanks to this cooperation, the team of MunicHMotorsport was able to very rapidly build up the required know-how in the field of Machine Learning in the 2017 season. In the 2018 season, we significantly improved the autonomous systems for vehicle control once again, so that the vehicle was able to successfully complete seven laps of the free track course at 17 km/h. This was the highest speed achieved in the first lap among all teams.

Artificial Intelligence for one day

In our one-day workshop “AI for one day” we aim to demystify AI together with you, make it tangible and comprehensible. It is also intended to give you a tool with which you can explain this topic to people not as familiar with IT and mathematics.

More about Machine Learning (in german)

Stefan Blum

Project Manager Artificial Intelligence and Machine Learning
+49 89 614551-0