Prof. Dr. Friedel Mager,
Vicepresident for Studies and Teaching, Rheinische Hochschule Köln
“Developer communities and global corporations are currently in a race for the performance of their AI models. As a university, it is important for us to be able to offer AI tools flexibly and across all manufacturers. Our students and research institutes also need a GDPR-compliant environment and security in the processing of their training data and conversations. On a high-performance platform with a highly up-to-date portfolio, as is the case with KISSKI. A remarkable service concept.”
Vicepresident for Studies and Teaching, Rheinische Hochschule Köln
“Developer communities and global corporations are currently in a race for the performance of their AI models. As a university, it is important for us to be able to offer AI tools flexibly and across all manufacturers. Our students and research institutes also need a GDPR-compliant environment and security in the processing of their training data and conversations. On a high-performance platform with a highly up-to-date portfolio, as is the case with KISSKI. A remarkable service concept.”
Prof. Dr. Cornelius Puschmann,
University of Bremen
“We use the KISSKI LLM API for automated content analysis, for example for analysing news articles from the Internet on reporting on climate change. The use of open source models is methodologically much cleaner than relying on non-transparent proprietary services and we also know what happens to the data. The API is very user-friendly and we can use more sophisticated models than would otherwise be available to us. In short, a great tool for our research
We use the project here, among other things:
Fabian Runge
Media Lab, Technische Universität Braunschweig
„The Application Laboratory for Artificial Intelligence and Big Data at the Federal Environment Agency (KI-Lab at UBA) makes methods of artificial intelligence practically usable for the authorities of the environment department. The KISSKI LLM service in particular supports us in this endeavour. Easy and data protection-compliant access to large language models enables us to find efficient solutions to our problems. One example of this is the creation of simple prototypes for the planned objection management system, which is to be used in the search for a final repository. We are delighted with the collaboration and look forward to further projects!’“
Media Lab, Technische Universität Braunschweig
„The Application Laboratory for Artificial Intelligence and Big Data at the Federal Environment Agency (KI-Lab at UBA) makes methods of artificial intelligence practically usable for the authorities of the environment department. The KISSKI LLM service in particular supports us in this endeavour. Easy and data protection-compliant access to large language models enables us to find efficient solutions to our problems. One example of this is the creation of simple prototypes for the planned objection management system, which is to be used in the search for a final repository. We are delighted with the collaboration and look forward to further projects!’“
Data Engineer,
AI-Lab Federal Environment Agency
“As the Media Lab at TU Braunschweig, we are constantly looking for solutions that support our work in the field of artificial intelligence for the university. The Chat-AI service from GWDG and KISSKI is proving to be very valuable in this context. Its availability via the Academic Cloud in Lower Saxony complements our internal tools at TU Braunschweig. An important aspect of Chat-AI is the data protection-friendly execution with local Large Language Models. The seamless integration of the OpenAI-compatible Chat-AI API into our existing open-source AI tools is an additional advantage. Chat-AI proves to be particularly reliable when demand is high, for example during AI training sessions. Another positive aspect is the comprehensive support, which helps us quickly with detailed questions.”
AI-Lab Federal Environment Agency
“As the Media Lab at TU Braunschweig, we are constantly looking for solutions that support our work in the field of artificial intelligence for the university. The Chat-AI service from GWDG and KISSKI is proving to be very valuable in this context. Its availability via the Academic Cloud in Lower Saxony complements our internal tools at TU Braunschweig. An important aspect of Chat-AI is the data protection-friendly execution with local Large Language Models. The seamless integration of the OpenAI-compatible Chat-AI API into our existing open-source AI tools is an additional advantage. Chat-AI proves to be particularly reliable when demand is high, for example during AI training sessions. Another positive aspect is the comprehensive support, which helps us quickly with detailed questions.”
Dr. Robert Haase,
Center for Scalable Data Analytics
and Artificial Intelligence Dresden / Leipzig
“My colleagues and I are very satisfied users of KISSKI’s ChatAI API. We use it in numerous research and training projects, typically in the context of code generation for [image] data analysis. The best open-weight model we have ever tested, LLAMA 3.1 405B, was only accessible to us via the KISSKI infrastructure. Furthermore, KISSKI has found its way into several open-source libraries that we offer to the life science community, where it is gratefully accepted as an alternative to commercial services whose privacy policies are not always completely transparent. It is also very important to us that we can use KISSKI’s ChatAI API for training to spread the targeted and responsible use of large language models in the community. Finally, I would like to thank the KISSKI team for their excellent service. I was in contact with my colleagues several times and always received feedback at the highest level within a very short time!”
Center for Scalable Data Analytics
and Artificial Intelligence Dresden / Leipzig
“My colleagues and I are very satisfied users of KISSKI’s ChatAI API. We use it in numerous research and training projects, typically in the context of code generation for [image] data analysis. The best open-weight model we have ever tested, LLAMA 3.1 405B, was only accessible to us via the KISSKI infrastructure. Furthermore, KISSKI has found its way into several open-source libraries that we offer to the life science community, where it is gratefully accepted as an alternative to commercial services whose privacy policies are not always completely transparent. It is also very important to us that we can use KISSKI’s ChatAI API for training to spread the targeted and responsible use of large language models in the community. Finally, I would like to thank the KISSKI team for their excellent service. I was in contact with my colleagues several times and always received feedback at the highest level within a very short time!”
Martin Stammann,
Transfer & Startup Hub Universität Göttingen
“We are very happy to use KISSKI’s services for our AI start-up teams. Both the support with computing power and the expert advice help the teams for the next steps on the start-up journey.”
Transfer & Startup Hub Universität Göttingen
“We are very happy to use KISSKI’s services for our AI start-up teams. Both the support with computing power and the expert advice help the teams for the next steps on the start-up journey.”
Data Lingua
“Consulting on complex topics such as the AI Act and medical device approval has helped us make significant progress. When we are established, we also want to use KISSKI’s computing resources to shorten our development times and get off to a better start. Great support and a great team!”
“Consulting on complex topics such as the AI Act and medical device approval has helped us make significant progress. When we are established, we also want to use KISSKI’s computing resources to shorten our development times and get off to a better start. Great support and a great team!”
Adrian Holt,
FloodWaive Predictive Intelligence GmbH
“At FloodWaive, we develop AI-supported hybrid models for fast, precise prediction and risk analysis of the effects of heavy rain and flood events. The cooperation with the German AI service centers, especially KISSKI, supports us significantly. Access to GPU-based HPC systems allows us to effectively further develop our deep learning models in research. The comprehensive support and resources provided have been instrumental in driving our scientific progress and technical innovation.”
FloodWaive Predictive Intelligence GmbH
“At FloodWaive, we develop AI-supported hybrid models for fast, precise prediction and risk analysis of the effects of heavy rain and flood events. The cooperation with the German AI service centers, especially KISSKI, supports us significantly. Access to GPU-based HPC systems allows us to effectively further develop our deep learning models in research. The comprehensive support and resources provided have been instrumental in driving our scientific progress and technical innovation.”
Manfred Steger,
Staatsinstitut für Frühpädagogik und Medienkompetenz (IFP)
“A lot of power without much knowledge of how it can be used usually ends in actionism and not in a meaningful application. There is a need for show and use cases for mainstream applications in different areas (e.g. upload R file and off you go, API LLM - tapped - off you go etc.), which can be tried out with 2-3 clicks.”
Staatsinstitut für Frühpädagogik und Medienkompetenz (IFP)
“A lot of power without much knowledge of how it can be used usually ends in actionism and not in a meaningful application. There is a need for show and use cases for mainstream applications in different areas (e.g. upload R file and off you go, API LLM - tapped - off you go etc.), which can be tried out with 2-3 clicks.”
