The German pharma company Boehringer Ingelheim has chosen to go with Kairntech.

Posted by Stefan Geißler on December 26, 2019

German Pharma Company Boehringer-Ingelheim licenses AI Platform from Kairntech.


Biberach, December 19th 2019 - The German pharma company Boehringer-Ingelheim (BI) and AI software startup Kairntech from Grenoble in France have announced their cooperation on tasks around the analysis of text-based information. The foundation of the cooperation is the licensing of Kairntech’s software platform Sherpa by Boehringer-Ingelheim.

"Drug Discovery is a knowledge–driven business", explains Dr. Karlheinz Spenny, Head of the Scientific Information Center at Boehringer's R&D site in Germany. "Much of what we know today resides as unstructured text in scientific publications, patents, conference abstracts or on websites. Making sense of (hidden) relations e.g. between genes and cellular targets. Such knowledge is crucial for discovery research scientists. Powerful AI and NLP (natural language processing) software is vital to accomplish knowledge discovery in an efficient and fast manner for our scientists."

Boehringer has therefore decided to license the Kairntech AI platform Sherpa. Kairntech and BI have assessed the performance of the software on a number of tasks such as document categorization and the recognition of domain specific entities in explorative projects with Boehringer-specific requirements. A key aspect of the platform is the fact that while powerful machine learning methods are implemented under the hood, the handling of the platform is kept simple and intuitive such that domain experts can train and tune new document processing models without having to engage in complex rule writing or programming.

"We are living in exciting times in the area of NLP and AI", says Stefan Geißler, Cofounder at Kairntech. "The field is progressing with relentless speed and use cases that would have required extensive manual efforts until recently, today can be addressed with appropriate methods such as Deep Learning and others."

Read the whole Press Release